Time Nick Message 03:57 MTDiscord my 2c: I already trust Buckaroo with my user data on blockexchange 04:01 MTDiscord on another note: Could we get https://github.com/minetest/minetest/pull/14260 ? It would be the cherry on top of the recent mapgen thread work 04:02 MTDiscord Currently adoption needed 09:51 sfan5 once we have a blog post ready someone should submit it to the subreddit 13:56 sfan5 rubenwardy: did you tweet about the release yet? 13:56 rubenwardy Waiting for a blog post 13:57 rubenwardy Which I believe luantic is working on? 13:58 sfan5 i see 13:58 sfan5 did someone update https://content.minetest.net/packages/Minetest/devtest/ and also the versions on cdb? 14:23 MTDiscord I am working on a blog post, yes 15:54 sfan5 please review #15388 so we can merge it before other work on the collision code happens 15:54 ShadowBot https://github.com/minetest/minetest/issues/15388 -- [no sq] Add unit tests for collisionMoveSimple by sfan5 16:53 ROllerozxa rubenwardy: what is the current rollout for 5.10.0 on google play, still 10%? 16:53 ROllerozxa I received an email from a teacher of a school in Scotland about how they had been using ROllertest for their students on Chromebooks since the newer versions of Minetest (5.9) had broken the controls there, but I had taken down ROllertest from Play store a day or so ago. I tested with Luanti 5.10.0 in an emulator and the mouse-keyboard control scheme now works fully when the touchscreen controls are disabled in the settings 16:53 ROllerozxa compared to 5.9, so it would be nice if they could get access to the new update 17:47 Krock sfan5: uploaded 5.10.0 devtest. https://content.luanti.org/packages/Minetest/devtest/releases/28405/edit/ someone please approve (just to be sure) 17:52 [MatrxMT] rollerozxa: it shouldn't even be necessary to change the settings, with 5.10.0 the kb/mouse controls should just work ("auto-toggling") 17:53 [MatrxMT] but perhaps that doesn't work in an emulator, if mouse input simulates touch input there 17:55 ROllerozxa I see, well I mentioned the setting just in case. but the chromebooks didn't have any touch support so it may default to keyboard and mouse controls then 18:01 Krock sfan5: https://github.com/minetest/minetest/pull/15388/files#diff-3f731fd96ee2702137e3c8e5736b1df7718db6e49e557c52afaaf7cb26f299bcR230 why 1.5 ? The object is located at (4, 1, 4), thus should collide with the nodes at y=0. Hence shouldn't it be y=0.5 ? 18:01 sfan5 accel is upwards 18:02 Krock oh indeed. Makes sense. Thanks. 18:02 sfan5 also even if it went down the cbox intentionally does not have that weird 0.5 offset 18:03 Krock this also means that pos_max_d is not respected? 18:03 sfan5 I have no idea what pos_max_d does 18:03 Krock hehe. it's an unused variable. 18:04 Krock you could as well set it to // TODO: removeme 8) 18:05 sfan5 🤦 18:07 Krock you might also like to use { } around individual tests to separate them better. That's just a personal preference, though. 18:09 MTDiscord Why not use Catch? These could be properly named SECTIONs then. 18:11 sfan5 the existing file does not use catch 18:12 Krock we kind of already have a test framework. just not as sophisticated as it would be possible with catch2 18:13 MTDiscord I know, I prefer catch2 18:13 Krock (by "test framework" I mean the test modules and the code around it for our unittests) 18:14 MTDiscord But I won't impose. Just jumped in to say that I think // frobnication \n { ... } is less nice than SECTION("frobnication") { ... }. 18:15 Krock fair enough 18:16 Krock > Frobnicate is a placeholder term in computer science used to represent a placeholder, undefined, or unknown action or functionality, often in documentation or code. 18:17 Krock interesting. kinda similar to Lorem Ipsum (text) and Foo Bar Baz (variables) 18:23 MTDiscord Krock: btw the changelog contains two minor errors. i would fix them myself but i don't have a wiki account. 18:24 MTDiscord "Improved collision detection at high dtime (lag) (kno10)" needs to be removed or annotated since it ended up being reverted. 18:24 MTDiscord The driver bug workaround should be attributed to sfan5 who first proposed it. 18:24 Krock anything else? 18:26 MTDiscord not anything i see right now 18:26 Krock thanks. corrected. 18:26 MTDiscord @celeron55 for the future: how do i best obtain a wiki account? 18:28 Krock https://forum.minetest.net/viewtopic.php?p=437468 should be updated to list the people with permissions 18:28 sfan5 i think wuzzy can create some 18:30 MTDiscord I will shoot Wuzzy a message on Matrix 18:46 sfan5 Krock: please re-check the PR 18:52 Krock > Temporary option for old move code 18:52 Krock so it has become permanent now? 18:56 sfan5 of course 18:57 Krock rereviewed 20:38 sfan5 thought: we could get some performance wins by rendering CAOs in a different order, grouped by same material 20:40 Krock CAOs aren't necessarily used in large quantities, but it could help a bit. 20:42 sfan5 uhh 20:42 sfan5 mobs?? 20:43 sfan5 different question: do we have test entities with transparency in devtest? 20:43 Krock I was more thinking about dropped items 20:43 Krock not that I'd know. 20:45 sfan5 hmm though actually mobs will not share the same VBOs, since animations are done in software 20:46 sfan5 🤷 20:46 SFENCE_arch The macOS build is in the notarization process. It probably went through the scripted part and it now waiting for a manual check from an Apple employer. 20:52 sfan5 nice 21:02 MTDiscord let's first reunfucktor skeletal animations ;) 21:04 sfan5 i lack experience and knowledge with graphics, but i don't think there are any low-hanging fruits to improve rendering performance in MT currently 21:04 sfan5 except 2d batching for GUIs but that's secondary 21:06 MTDiscord moving skeletal animation to the GPU isn't very low-hanging but it is doable 21:08 sfan5 will that improve performance? it's not like doing some trigonometry on the cpu is slow 21:10 MTDiscord sfan5: it's a lot of linear algebra you're doing on the CPU, you have to transform every vertex. for a hotfix optimizing the CPU skinning manually might be an option but we really want to be doing this on the GPU. 21:10 MTDiscord see e.g. #13043 21:10 ShadowBot https://github.com/minetest/minetest/issues/13043 -- CSkinnedMesh.* consumes a lot of CPU (with many mobs) 21:10 sfan5 ah I see 22:22 MTDiscord I've written a first draft for the release blog post: blog#154 22:22 pgimeno https://github.com/minetest/blog/pull/154 -- 5.10.0 release blog post first draft by appgurueu 22:23 MTDiscord Thanks pgimeno-bot 22:28 MTDiscord I will take a look at it in a bit 22:28 MTDiscord Thanks 23:22 MTDiscord I've requested to join the minetestdevs team on launchpad. Do we now want to set up another team called luantidevs given the rename?