| Time |
Nick |
Message |
| 00:43 |
|
amfl2 joined #minetest |
| 01:08 |
ireallyhateirc |
I can't load a texture from another mod for some reason. I have a "pcity_nodes" mod that uses the texture and "pcity_common" that stores the texture in the "textures" directory |
| 01:09 |
ireallyhateirc |
Could not load image "/home/user/.minetest/games/perfect-city/mods/pcity_common/textures/pcity_common_palette.png" while building texture; Creating a dummy image |
| 01:09 |
ireallyhateirc |
but the path itself exists and is correct |
| 01:09 |
ireallyhateirc |
I made sure to make pcity_common a dependency of pcity_nodes |
| 01:10 |
ireallyhateirc |
Any ideas why it doesn't work? |
| 01:13 |
ireallyhateirc |
Ok, never mind. I used file path for tiles instead of simply the file name |
| 01:41 |
|
v-rob joined #minetest |
| 02:21 |
|
IamPyu joined #minetest |
| 03:25 |
|
calculon joined #minetest |
| 03:27 |
|
SFENCE_ joined #minetest |
| 03:52 |
silverwolf73828 |
I love the new textures and shaders! The ray tracing and the sun reflecting on the water make everything look so real! |
| 04:07 |
|
sparky4 joined #minetest |
| 04:37 |
|
v-rob joined #minetest |
| 04:44 |
|
chilledfrogs joined #minetest |
| 05:00 |
|
MTDiscord joined #minetest |
| 05:35 |
|
gregon joined #minetest |
| 06:02 |
|
lagash joined #minetest |
| 06:22 |
|
Verticen joined #minetest |
| 06:38 |
|
Baytuch joined #minetest |
| 06:42 |
|
lagash joined #minetest |
| 06:52 |
|
SFENCE joined #minetest |
| 07:30 |
|
cranezhou joined #minetest |
| 07:49 |
|
tarsovbak joined #minetest |
| 08:46 |
|
jaca122 joined #minetest |
| 08:53 |
|
Talkless joined #minetest |
| 09:02 |
|
Talkless joined #minetest |
| 09:22 |
|
Warr1024 joined #minetest |
| 09:24 |
|
gregon joined #minetest |
| 09:25 |
|
tarsovbak joined #minetest |
| 09:45 |
|
gregon joined #minetest |
| 09:46 |
|
Warr1024 joined #minetest |
| 09:52 |
|
MacroFaxSax joined #minetest |
| 10:08 |
|
TomTom joined #minetest |
| 10:29 |
|
Glaedr joined #minetest |
| 10:39 |
|
jluc joined #minetest |
| 10:45 |
|
gregon joined #minetest |
| 10:45 |
|
ireallyhateirc joined #minetest |
| 11:43 |
|
fluxionary joined #minetest |
| 11:43 |
|
tarsovbak joined #minetest |
| 12:08 |
|
tarsovbak joined #minetest |
| 12:09 |
|
gregon joined #minetest |
| 12:28 |
|
yezgromafic joined #minetest |
| 13:06 |
|
Baytuch joined #minetest |
| 13:25 |
|
tarsovbak joined #minetest |
| 13:36 |
MTDiscord |
<_devsh_> @luatic there is not saving from rgb8 instead of rgba8, no GPU supports it I'd add support for S3TC (dxt/bc1-3) formats if you want faster rendering or upload |
| 13:38 |
MTDiscord |
<_devsh_> Sfan5 making CImage a template is a mistake, I'd also focus on actually supporting more formats transparently, we support all 150 vulkan formats without having to worry much |
| 13:43 |
sfan5 |
I'm sure that'd be nice but I don't see an use case for working with image formats other than rgba8 on the cpu |
| 13:44 |
MTDiscord |
<_devsh_> I mean you have to load them from files somehow? |
| 13:45 |
sfan5 |
either the API gives us rgba8 or we have to convert it once |
| 13:45 |
MTDiscord |
<_devsh_> You're probably leaving quite a bit of FPS on the floor by not using dxt and astc compression |
| 13:46 |
sfan5 |
they're faster to render? |
| 13:46 |
MTDiscord |
<_devsh_> R8 and RG textures are also somewhat useful |
| 13:47 |
MTDiscord |
<_devsh_> Yes, they're faster esp on Integrated and mobile GPUs |
| 13:47 |
MTDiscord |
<_devsh_> There's a nice library and format called Basis Universal |
| 13:47 |
MTDiscord |
<luatic> Okay, then ARGB8 it is |
| 13:47 |
MTDiscord |
<_devsh_> The texture on disk is similar side to jpeg but not as lossy |
| 13:48 |
MTDiscord |
<_devsh_> And you tell basis-u to transcode into S3TC or ASTC or ETC when loading |
| 13:49 |
sfan5 |
sounds worth investigating |
| 13:49 |
sfan5 |
but my intuition says we should fix our other render problems first |
| 13:49 |
MTDiscord |
<_devsh_> Reason is that desktops (including integrated GPU support only s3tc or ASTC), apple silicon/iOS supports ETC and ASTC, and mobile ASTC |
| 13:50 |
MTDiscord |
<_devsh_> You're digging yourself into a bigger hole by not handling formats properly and hard coding everything into one |
| 13:50 |
MTDiscord |
<_devsh_> Also what you call is actually argb_srgb |
| 13:51 |
MTDiscord |
<luatic> Yeah it is not linearized |
| 13:51 |
|
vampirefrog joined #minetest |
| 13:52 |
MTDiscord |
<luatic> I don't think that having every method switch on color format is the right way to implement multiple color formats, even if we do end up wanting them, though |
| 13:53 |
MTDiscord |
<_devsh_> Depends what your methods do |
| 13:55 |
MTDiscord |
<_devsh_> I have just templated block encode and decode functions, then you can generate a switch on lambdas for every case of a compound function |
| 13:58 |
|
tarsovbak joined #minetest |
| 13:59 |
|
SFENCE joined #minetest |
| 14:02 |
|
tarsovbak1 joined #minetest |
| 14:06 |
|
tarsovbak joined #minetest |
| 14:19 |
|
gregon joined #minetest |
| 14:26 |
MTDiscord |
<_devsh_> Sfan5 https://realtimecollisiondetection.net/blog/?p=86 |
| 14:26 |
MTDiscord |
<_devsh_> Rewrite onRender |
| 14:26 |
MTDiscord |
<_devsh_> Everything downstream of frustum culling |
| 14:27 |
MTDiscord |
<_devsh_> And instead of traversing whole scene node and masking out stuff to draw, traverse once and bucket the mesh buffer & drawcall metadata onto separate vectors that you'll later sort |
| 14:32 |
MTDiscord |
<_devsh_> Below 64k entries use std::sort |
| 14:32 |
MTDiscord |
<_devsh_> Above grab yourself a radix sort |
| 14:44 |
sfan5 |
thanks for the link |
| 14:52 |
|
silverwolf73828 joined #minetest |
| 15:05 |
|
iisu joined #minetest |
| 15:06 |
iisu |
Hi, I can't log in to forum. |
| 15:06 |
iisu |
It says: "You have exceeded the maximum number of login attempts allowed. In addition to your username and password the invisible reCAPTCHA v3 will be used to authenticate your session." |
| 15:07 |
iisu |
Well, I don't see any invisible CAPTCHA. |
| 15:09 |
sfan5 |
it's invisible, that's the point |
| 15:10 |
sfan5 |
you don't need to interact with it |
| 15:12 |
iisu |
So what do I do? |
| 15:14 |
sfan5 |
enter your username and password, click login |
| 15:14 |
iisu |
Doesn't work. |
| 15:14 |
iisu |
It shows that message again. |
| 15:16 |
sfan5 |
then you got the password wrong |
| 15:17 |
ROllerozxa |
the "invisible" recaptcha is visible in the bottom right corner. if it is suspicous of you then it will pop up a challenge when you press the form submit button, otherwise it will let you pass. if it doesn't work beyond that then you typed the wrong password |
| 15:19 |
iisu |
I did get the password wrong in the beginning, which I assume is why I'm getting this message in the first place. But now I'm sure I got it right. |
| 15:20 |
iisu |
I tried resetting my password so maybe that's why, idk. |
| 15:20 |
iisu |
And password reset doesn't work too. It says: "The requested page could not be found.". |
| 15:21 |
MTDiscord |
<jordan4ibanez> https://tenor.com/view/soccer-robot-goalie-gif-23578122 |
| 15:21 |
ROllerozxa |
yeah password resets are broken, you will need to contact a forum staff member yourself to get your password changed |
| 15:21 |
MTDiscord |
<jordan4ibanez> That's exactly what a robot would say :squint: |
| 15:24 |
iisu |
ok, so you are aware of the broken password resets. |
| 15:25 |
|
Alias joined #minetest |
| 15:25 |
iisu |
Is my old password no longer valid because I clicked on reset password, or is this an issue with the CAPTCHA? |
| 15:25 |
sfan5 |
what's the name of your account? |
| 15:26 |
iisu |
iisu |
| 15:28 |
sfan5 |
i dont see any indication that your account is locked or similar |
| 15:29 |
ROllerozxa |
I just tested logging into my forum account with an incorrect password in an incognito window until I got the recaptcha verification and then tried logging in again with my correct password and it was successful |
| 15:30 |
sfan5 |
iisu: can I send you the new password by email? |
| 15:30 |
iisu |
ok |
| 15:31 |
sfan5 |
done |
| 15:33 |
iisu |
I'm still getting the message. |
| 15:33 |
sfan5 |
any browser extensions that may interfere with recaptcha? |
| 15:35 |
iisu |
ok, I see the CAPTCHA now. |
| 15:35 |
iisu |
It was Privacy Badger. |
| 15:36 |
iisu |
I logged in. |
| 15:36 |
iisu |
Thanks. |
| 15:40 |
ireallyhateirc |
Is there anything extra I need to do in order to have a mesh with 2 materials? I mapped 2 materials on the mesh to 2 textures, added 2 tiles in the nodedef but it still uses only the first tile. |
| 15:41 |
ireallyhateirc |
strangely enough when I break the node I get particles from the second tile too |
| 15:56 |
|
DarkLordofDebian joined #minetest |
| 15:58 |
|
MacroFaxSax joined #minetest |
| 16:14 |
ireallyhateirc |
Okay, I managed to find the answer on the forums |
| 16:14 |
sfan5 |
out of interest, what is it? |
| 16:18 |
ireallyhateirc |
a tap/faucet model |
| 16:18 |
|
Lunatrius` joined #minetest |
| 16:18 |
ireallyhateirc |
https://files.catbox.moe/ti044z.png |
| 16:19 |
sfan5 |
the answer, I meant |
| 16:19 |
ireallyhateirc |
https://forum.luanti.org/viewtopic.php?t=22990 |
| 16:19 |
ireallyhateirc |
basically selecting "Material groups" checkbox while exporting |
| 16:19 |
MTDiscord |
<luatic> ah so you're using obj, good |
| 16:19 |
MTDiscord |
<luatic> because in gltf i should have done this properly |
| 16:20 |
ireallyhateirc |
should I try? |
| 16:20 |
MTDiscord |
<luatic> if you want to, sure |
| 16:21 |
ireallyhateirc |
I'm not sure if the solution is optimal, I have 2 tap models one general metallic texture and one color palette |
| 16:22 |
ireallyhateirc |
the color palette is 5KB and I'd like to use it for nearly all flat single-color surfaces |
| 16:22 |
MTDiscord |
<luatic> It'd probably be better for performance to have it all be one mesh buffer but I guess you're not gonna have that many faucets |
| 16:23 |
ireallyhateirc |
but I heard using multiple materials for a mesh decreases performance |
| 16:23 |
ireallyhateirc |
every single flat is going to have 2-3 faucets |
| 16:23 |
ireallyhateirc |
I'll have 3-5 faucet types in the game |
| 16:24 |
ireallyhateirc |
an apartment building would be 5 flats times 13 floors so about 195 faucets per building |
| 16:24 |
MTDiscord |
<luatic> ireallyhateirc: yeah, as i said, multiple materials for a mesh does tend to decrease performance, because each of them basically needs to be rendered separately (though i think the mapblock meshgen should do material based batching so you could be fine there) |
| 16:24 |
ireallyhateirc |
but Luanti should be smart enough not to render them all? |
| 16:25 |
ireallyhateirc |
using the color palette saves time when creating textures and saves disk space |
| 16:26 |
MTDiscord |
<_devsh_> @luatic you don't need geometry shader to billboard particles, you can do it in vertex shader if you divide gl_VertexID by 4 to get your particle ID |
| 16:27 |
MTDiscord |
<_devsh_> And modulo 4 to get the local quad vertex id |
| 16:27 |
MTDiscord |
<_devsh_> P.s. if you want high particle counts and no CPU lag, look into ECS |
| 16:27 |
MTDiscord |
<luatic> I know that I don't need the geometry shader to billboard particles |
| 16:28 |
MTDiscord |
<_devsh_> So what's the talk about geo shaders |
| 16:29 |
MTDiscord |
<_devsh_> Also another reason to do programmable pulling from a buffer texture, because you don't need to duplicate particle data in each vertex |
| 16:29 |
MTDiscord |
<luatic> I think sfan was thinking of using the geo shader to do two things: One, turn a single position into a quad, two, rotate this quad properly. |
| 16:29 |
MTDiscord |
<luatic> I think we can directly skip the latter if we just transform less in the vertex shader, and I don't care too much about the former yet. |
| 16:29 |
MTDiscord |
<_devsh_> You can do both in vertex |
| 16:30 |
MTDiscord |
<_devsh_> Geo shader was a mistake, that's why mobile GPUs don't have them |
| 16:30 |
sfan5 |
you can invent new vertices in a vertex shader or how does that work? |
| 16:30 |
MTDiscord |
<_devsh_> No, you always draw 4x more verts |
| 16:30 |
sfan5 |
ah I see |
| 16:31 |
MTDiscord |
<_devsh_> But you don't use hardware vertex attribute inputs |
| 16:31 |
MTDiscord |
<_devsh_> You have a vertex format that has 0 bytes |
| 16:31 |
MTDiscord |
<_devsh_> 0 slots used |
| 16:31 |
MTDiscord |
<_devsh_> And you get your particle data from SSBO or Buffer Texture |
| 16:32 |
MTDiscord |
<_devsh_> https://ktstephano.github.io/rendering/opengl/prog_vtx_pulling |
| 16:33 |
MTDiscord |
<_devsh_> You can pull at a lower rate |
| 16:40 |
|
gregon1 joined #minetest |
| 16:42 |
rubenwardy |
Hi everyone, Luanti is going to be at FOSDEM 25 in Brussels! Come along on 1st-2nd February for stickers and more https://fosdem.org/2025/ |
| 16:48 |
|
SFENCE joined #minetest |
| 16:50 |
ireallyhateirc |
behold the taps: https://files.catbox.moe/10sbr6.png |
| 16:51 |
MTDiscord |
<luatic> faucet_game |
| 16:54 |
ireallyhateirc |
I'll make a monster out of it later |
| 17:01 |
Krock |
sonic screwdriver! |
| 17:01 |
Krock |
bzzzzz bzzzzz |
| 17:01 |
MTDiscord |
<et086> im pretty sure thats the omnidriver https://github.com/OgelGames/omnidriver |
| 17:02 |
ireallyhateirc |
yeah it's the omnidriver |
| 17:02 |
Krock |
omimus prime |
| 17:26 |
|
BuckarooBanzai joined #minetest |
| 17:27 |
|
BuckarooBanzai joined #minetest |
| 17:29 |
|
BuckarooBanzai joined #minetest |
| 17:34 |
|
BuckarooBanzai joined #minetest |
| 17:43 |
|
fbievan joined #minetest |
| 18:09 |
|
tarsovbak joined #minetest |
| 18:15 |
ireallyhateirc |
let that sink in https://files.catbox.moe/gf5hbc.png |
| 18:18 |
ireallyhateirc |
Are there any unforseen consequences of making meshes that are bigger than ordinary nodes? |
| 18:21 |
MTDiscord |
<zmv7> Your sink and water tap are single node? |
| 18:21 |
Krock |
they might intersect with other nodes but that's about it |
| 18:21 |
Krock |
problems arise when you have collision boxes and selection boxes that are oversized |
| 18:22 |
ireallyhateirc |
zmv7, no the sink is separate from the tap |
| 18:22 |
MTDiscord |
<zmv7> Ah. Single-node sink+water tap will give able to place the mirror at the nice height |
| 18:23 |
ireallyhateirc |
I use a different player scale than minetest_game. Here the nodes are ~0.55 meters while in MTG it's 1 meter |
| 18:23 |
ireallyhateirc |
similar trick to what backroom test uses |
| 18:24 |
MTDiscord |
<zmv7> It solves these things then |
| 18:24 |
ireallyhateirc |
to fit a mirror I'd have to move both the sink and the tap slightly down |
| 18:25 |
ireallyhateirc |
but that's not a big deal |
| 18:29 |
ireallyhateirc |
It's too high anyway so I'll correct it now |
| 18:38 |
ireallyhateirc |
Now time for a toilet |
| 18:41 |
|
jluc joined #minetest |
| 18:42 |
|
fbievan joined #minetest |
| 18:56 |
|
SFENCE_arch joined #minetest |
| 19:17 |
|
SFENCE joined #minetest |
| 19:54 |
|
SFENCE joined #minetest |
| 19:57 |
|
Verticen joined #minetest |
| 19:59 |
|
SFENCE joined #minetest |
| 20:02 |
|
Helenah joined #minetest |
| 20:05 |
Helenah |
Hi, I need some help with modeling for Minetest, so what I'm basically wanting to know is, how big is a full node in real-world terms? To be more specific and precise on what exactly I'm trying to do, I've in OpenSCAD created a triangle, I intend the triangle to be 3D, so it will therefore need a thickness to it, I intend to make it a sign worth in thickness but obviously, my issue here is, how |
| 20:05 |
Helenah |
thick? Because I do not know the scale of things, I've scored your wiki, I've scored your forums for this information, it doesn't help that it takes 20 days for these two sites to load for whatever reason that may be, I found no information. |
| 20:05 |
|
SFENCE joined #minetest |
| 20:06 |
Helenah |
To be honest, I originally had the impression that your web services are down, turns out they're not though, could be a bad node between me and the services, or it could be you guys. |
| 20:06 |
Krock |
normal day |
| 20:06 |
Krock |
the servers are tired, that's all. |
| 20:07 |
Krock |
maybe relevant: https://forum.luanti.org/viewtopic.php?p=440597 |
| 20:07 |
Krock |
!title |
| 20:07 |
MinetestBot |
Krock: [Animated model] How do I export a working gltf/glb from my blends? - Luanti Forums |
| 20:07 |
Krock |
> scale is off by ~10... no.. x0.316 to get about the original size |
| 20:09 |
Helenah |
Wait, exporting from Blender? I use OpenSCAD, and to be honest, the software I'm using is not the issue, as said in my question, I do not know what a full node is in real terms. |
| 20:10 |
Krock |
I'd suggest try & error until it fits |
| 20:10 |
Krock |
have a cube of 1x1x1 meter and adjust it as needed |
| 20:10 |
Helenah |
Thank, I was hoping to avoid that, I guess that'll have to do then. |
| 20:10 |
Helenah |
*Thanks |
| 20:10 |
Helenah |
Wait... |
| 20:11 |
Helenah |
You just answered my question. |
| 20:11 |
Helenah |
So a full node is 1x1x1m? |
| 20:11 |
Krock |
I'd personally say that it's 1 meter but there's some people that argue it's not |
| 20:12 |
Krock |
After all it all depends on how the model is saved to the disk and then read back as vertices |
| 20:12 |
Krock |
and I really don't know what the scaling factor is in that case |
| 20:12 |
ireallyhateirc |
Helenah, in Blender one node = 1 meter |
| 20:12 |
Helenah |
Okay, well I guess in that case... 1 x 1 x 1 m is a good starting point, thanks. |
| 20:12 |
ireallyhateirc |
so making a 1 m cube will do the thing |
| 20:12 |
Helenah |
Yes, that should be the case in OpenSCAD too, I'll do it now, thanks. |
| 20:13 |
Helenah |
That's providing Blender is to scale, of course. :) |
| 20:13 |
Krock |
people are using Blender for all sorts of things nowadays, so I'd be surprised if it's wrong. |
| 20:15 |
|
v-rob joined #minetest |
| 20:17 |
sfan5 |
can blender slice my vegetables |
| 20:17 |
Krock |
if you 3D-print a vegetable slicing machine, then yes. |
| 20:18 |
Krock |
or find a way to abuse the scripting engine to control an Arduino board to spin a knife real fast |
| 20:26 |
|
SFENCE joined #minetest |
| 20:30 |
|
___nick___ joined #minetest |
| 20:33 |
|
lmat joined #minetest |
| 20:33 |
|
mrkubax10 joined #minetest |
| 20:35 |
|
SFENCE joined #minetest |
| 20:40 |
|
kattsand joined #minetest |
| 20:42 |
ireallyhateirc |
Behold the Seat: https://files.catbox.moe/as7nuf.png |
| 20:48 |
|
gregon joined #minetest |
| 20:58 |
|
kattsand left #minetest |
| 21:02 |
|
Galisma joined #minetest |
| 21:04 |
[MatrxMT] |
<🇬regon> Type "sudo apt install blender && LIBGL_ALWAYS_SOFTWARE=1 blender" in terminal if you are on Debian. This command will work on any videocard (and tested on Blender 2.82) |
| 21:05 |
[MatrxMT] |
<🇬regon> @Helenah Type "sudo apt install blender && LIBGL_ALWAYS_SOFTWARE=1 blender" in terminal if you are on Debian. This command will work on any videocard (and tested on Blender 2.82) |
| 21:06 |
gregon |
Oh, it sended message two times |
| 21:15 |
|
___nick___ joined #minetest |
| 21:26 |
|
___nick___ joined #minetest |
| 21:30 |
|
erle joined #minetest |
| 21:32 |
|
___nick___ joined #minetest |
| 21:33 |
Helenah |
[MatrxMT]: Why are you highlighting me? |
| 21:34 |
Helenah |
I don't understand the relevance. |
| 21:42 |
|
Thermoriax joined #minetest |
| 21:53 |
|
vampirefrog joined #minetest |
| 22:48 |
|
Galisma joined #minetest |
| 23:04 |
MinetestBot |
[git] HybridDog -> minetest/minetest: Remove unused and rarely used irrlicht color functions (#15442) 7295b6c https://github.com/minetest/minetest/commit/7295b6c88cbadb188871162d4d76e7c50faef9d1 (2024-11-17T23:02:53Z) |
| 23:04 |
MinetestBot |
[git] veprogames -> minetest/minetest: Replace occurences of 'wiki.minetest.net' with 'wiki.luanti.org' a8ea165 https://github.com/minetest/minetest/commit/a8ea1650425f83a7573314effae061823ff52de8 (2024-11-17T23:04:32Z) |
| 23:14 |
|
Thomas-S joined #minetest |
| 23:14 |
|
Thomas-S joined #minetest |
| 23:35 |
|
panwolfram joined #minetest |
| 23:59 |
|
jluc_ joined #minetest |
