| Time |
Nick |
Message |
| 00:05 |
|
turtleman joined #luanti-dev |
| 01:05 |
|
Eragon joined #luanti-dev |
| 02:08 |
|
nekobit joined #luanti-dev |
| 03:28 |
|
sfence_ joined #luanti-dev |
| 04:00 |
|
MTDiscord joined #luanti-dev |
| 07:10 |
|
m42uko_ joined #luanti-dev |
| 08:00 |
|
YuGiOhJCJ joined #luanti-dev |
| 09:33 |
|
YuGiOhJCJ joined #luanti-dev |
| 09:44 |
|
Juri2 joined #luanti-dev |
| 09:44 |
|
Niklp1 joined #luanti-dev |
| 11:30 |
|
turtleman joined #luanti-dev |
| 11:49 |
sfan5 |
it's great how all the "oh we can just do it like this, it will be fine" decisions from the past fall on my feet to fix with the array texture PR |
| 12:18 |
sfan5 |
the way the shadow map is rendered is basically a hack. it replaces all node materials with a single shader |
| 12:18 |
sfan5 |
predictably that shader does not handle different alpha modes, waving or other node specialities I can't think of |
| 12:33 |
sfan5 |
the deeper you dig there appear to be very few parts of Luanti code that aren't half-baked |
| 12:33 |
sfan5 |
no serious game dev would want to use an engine where even the clearly documented behavior is wrong |
| 12:59 |
|
sfence_ joined #luanti-dev |
| 14:10 |
sfan5 |
ready for review: #16574 |
| 14:10 |
ShadowBot |
https://github.com/luanti-org/luanti/issues/16574 -- [no squash] Introduce array textures for node rendering by sfan5 |
| 15:12 |
|
citrons joined #luanti-dev |
| 15:20 |
|
SFENCE joined #luanti-dev |
| 16:31 |
Krock |
at least we have comments |
| 16:31 |
Krock |
and thanks to you there's a limit to how much garbage builds up in the engine over the years |
| 17:31 |
sfan5 |
unfortunately I do not have unlimited time and motivation |
| 17:42 |
Krock |
Trivia: we have at least 3 linear interpolation implementations: numericalBlend getLerp linearInterpolation |
| 17:49 |
|
SFENCE joined #luanti-dev |
| 17:56 |
|
SFENCE joined #luanti-dev |
| 18:12 |
MTDiscord |
<josiah_wi> C++20 has a linear interpolation in the standard library. |
| 18:13 |
MTDiscord |
<josiah_wi> (std::lerp) |
| 18:41 |
|
SFENCE joined #luanti-dev |
| 18:42 |
MTDiscord |
<herowl> We're at C++17 for now, right? |
| 18:44 |
MTDiscord |
<luatic> irrlicht v3f's aren't arithmetic types though, right |
| 18:59 |
sfan5 |
fact: of Luanti's install basis on Google Play so few people have GLES 2.0 that they don't even show up in stats |
| 18:59 |
|
SFENCE joined #luanti-dev |
| 19:02 |
Krock |
huh? didn't we remove GLES 1.0 support? What are they using instead? |
| 19:02 |
sfan5 |
put differently, practically nobody doesn't have GLES 3.x |
| 19:02 |
Noisytoot |
untrue |
| 19:03 |
Noisytoot |
both my pinephone and thinkpad x200 lack GLES 3.x |
| 19:03 |
sfan5 |
your pinephone and thinkpad are not in the google play statistics |
| 19:03 |
|
jstein joined #luanti-dev |
| 19:04 |
sfan5 |
anyway where I am going with this is not that we're going to raise the minimum version, but rather that if the array texture code is fixed to work on GLES practically every mobile user will be able to profit from it |
| 19:05 |
user333_ |
what are the differences between OpenGL and (Open?)GLES? I know that GLES is for mobile devices, but is there any major difference other than that? |
| 19:08 |
Krock |
GLES only has a subset of the commands for simplicity |
| 19:08 |
sfan5 |
the main part is that it has additional restrictions but specifically for 3.x they also got rid of lots of legacy compatibility stuff that the respective OpenGL version still has |
| 19:09 |
user333_ |
i assumed it was a cut-down version of standard OpenGL, looks like i was right |
| 21:15 |
|
SFENCE joined #luanti-dev |
| 21:34 |
|
SFENCE joined #luanti-dev |
| 22:33 |
|
panwolfram joined #luanti-dev |
