Time Nick Message 01:10 kilbith if not, then the "release early / release often" model cannot scale on MT because it effectively makes us lose weeks of productivity during the freeze 08:44 MTDiscord why is anti-aliasing off by default? 08:44 MTDiscord it makes the game look like shit 99.9% of the time in my experience 08:44 MTDiscord every laptop ive had anyway 08:45 MTDiscord better question, why the fuck doesnt it work the way it should either 08:45 MTDiscord you still see the sharp pixelated edges 08:48 MTDiscord "you still see the sharp pixelated edges" 16x textures tend to do that lol 08:53 MTDiscord no I mean... I can see the pixels on my screen 09:41 MTDiscord bruh 09:41 MTDiscord antialiasing can't exactly magically increase the pixel density of your monitor 09:41 MTDiscord you need supersampling to get rid of those edges 09:42 MTDiscord you could tell GL to render the triangles antialiased 09:42 MTDiscord but that comes with blending issues around the edges IIRC 09:42 MTDiscord mmye 09:42 MTDiscord anyways @FatalError, once again seasoned, the antialiasing you're using is a texture interpolation only thing 09:42 MTDiscord and the antialiasing you'd have to use for smooth edges - MSAA / supersampling - is buggy 09:42 MTDiscord I wonder how irrlicht does the antialiasing? the minetest antialiasing is basically just telling irrlicht to do x amount of AA 09:42 MTDiscord ~cdb gop 09:43 MTDiscord ~cdb go 09:43 MTDiscord this is minetest-dev, I don't think that works here 09:43 MTDiscord oof 09:43 MTDiscord anyways, we're spamming -dev again 09:43 MTDiscord sorry 09:44 MTDiscord Anyways, here's what I wanted to quote from the Go Readme: > Smooth GUI image scaling should be enabled (gui_scaling_filter = true) > Enable trilinear texture filtering (trilinear_filter = true), mipmapping (mip_map = true) & anisotropic filtering (anisotropic_filter = true) for smoother board appearance > Consider enabling multi-sample anti-aliasing (msaa = 2) for smooth board edges if it doesn't trigger a nasty rendering bug on your 09:44 MTDiscord setup 09:44 MTDiscord Nasty rendering bug is https://github.com/minetest/minetest/issues/9072 09:45 sfan5 the setting is called `fsaa` 09:45 MTDiscord Indeed it is, thanks, will correct 09:46 MTDiscord I've had it disabled anyways since I have the nasty rendering bug ;) 09:46 sfan5 "I wonder how irrlicht does the antialiasing" it just tells the driver to do it as described in GL_ARB_multisample 10:56 schwarzwald[m] Who insured that `util/string.h` won't be mistaken for the `` STL header? 11:02 schwarzwald[m] s/insured/ensured/ 11:40 sfan5 that's probably always included via 11:41 sfan5 and/or perhaps the util folder is never directly added to the include path 11:41 schwarzwald[m] I think the latter is correct. Seems risky, though. 11:43 sfan5 the only #include outside of lib/ is in src/util/sha256.c 11:43 schwarzwald[m] There are also some includes in Irrlicht. 11:44 schwarzwald[m] And all includes technically include it indirectly as well. 11:44 schwarzwald[m] Maybe we could rename the file to `string_ext.h` or something just to be safe. 11:47 sfan5 is ensured to always include the system header 11:48 sfan5 or at least libstdc++ does, no idea if the standard mandates this 12:32 rubenwardy <> will include SYSTEM includes first 12:34 sfan5 IIRC it wasn't that simple, search the issue tracker for "event.h" 13:04 erle schwarzwald[m] are you saying this a problem? https://mister-muffin.de/p/rfx6.png 13:06 erle like, should there be no relation between src/util/sha256.o and /usr/include/string.h? 13:07 erle i do believe that having a single C file in the source code is a bit weird 13:16 schwarzwald[m] I do not fully understand why, but will include the wrong header if the util directory is added directly to the include paths. I haven't been able to produce a minimum reproducible example yet. 13:19 rubenwardy you shouldn't add the util directly to the include paths 13:19 schwarzwald[m] Why not? 13:29 schwarzwald[m] I will comply anyway, but I'm curious what the reasons are. 13:32 rubenwardy same reason you don't add other directories to the include path 13:34 rubenwardy it blurs the directories together 13:34 schwarzwald[m] Ah, so we keep the include paths prefixed in the src dir so it's clear where the includes are coming from? 16:50 sfan5 yes 19:04 fluxionary does anyone know if there's an existing issue about the fact that mouse-click events are triggered twice, both when pressing and releasing the button? 19:05 sfan5 doubt it 19:05 MTDiscord Mac? 19:06 fluxionary alright, i'll try to make one and document what i understand as best as i can 19:07 fluxionary LandarVargan: is that a question to me? no, linux, though i think the issue is probably cross-platform. 19:09 erle fluxionary include repro instructions that are as exact as possible, i'll try on linux 19:10 fluxionary erle: i might have found some related issues 19:10 fluxionary the most common way this manifests is when a player tries to drag something from one inventory to another, but the wrong item ends up in the other inventory (and the intended item ends up replacing the wrong item in the first inventory) 19:11 erle i have indeed encountered that bug! 19:11 fluxionary you can repro just by moving the mouse while click-picking up something. 19:11 erle but i do not see the connection 19:11 erle and never filed it (mea maxima culpa) 19:12 fluxionary https://github.com/minetest/minetest/issues/4477 seems to be describing the same thing, at least in part 19:13 sfan5 that sounds like a focus mistake, not what you're describing 19:14 fluxionary yeah, i think i'm gonna make a new issue and include repro instructions and a video 19:15 erle fluxionary FYI: the inventory bug happens more often when dragging under lag conditions. since inventory drawing is for some reason extremely heavy, it is more likely to occur the more things you see. you can try that in devtest i think. 19:16 erle client lag conditions i mean (rendering lag), not network lag or server lag 19:16 fluxionary seems reasonable. i bet i can replicate it by manually setting a lower max frame rate tho 19:18 fluxionary actually i can replicate it very easily w/out lag 19:40 fluxionary https://github.com/minetest/minetest/issues/12596