| Time |
Nick |
Message |
| 00:38 |
MTDiscord |
<wsor4035> ruben: since the website has migrated(cant just tweak previous pr), anything new for website version bump, or just adjust stuff in new location? |
| 01:13 |
rubenwardy |
it's the same - stuff in the data dir, and stuff in _includes/pages/downloads |
| 02:30 |
|
mrcheese joined #luanti-dev |
| 03:53 |
|
YuGiOhJCJ joined #luanti-dev |
| 05:00 |
|
MTDiscord joined #luanti-dev |
| 05:09 |
|
fluxionary joined #luanti-dev |
| 06:06 |
|
fluxionary joined #luanti-dev |
| 06:44 |
|
fluxionary joined #luanti-dev |
| 08:40 |
rubenwardy |
Doesn't 5.15 come with texture arrays? If there's a significant performance improvement that should be on the highlights |
| 08:42 |
sfan5 |
yes |
| 08:51 |
sfan5 |
I edited the forum post a bit |
| 08:53 |
rubenwardy |
For the blog post, do you have any figures on the improvement? |
| 08:54 |
rubenwardy |
Features? Anything other than array textures? |
| 09:01 |
sfan5 |
people have reported 2x to 10x in the PR, could use that with an "up to" to be cautious |
| 09:07 |
sfan5 |
looking through the changelog: "Add inventory image animation API" deserves a mention, dynamic shadows on Android, maybe non-cubic chunk size |
| 09:07 |
sfan5 |
mainmenu features: dark mode, ESC dialog, taking screenshots |
| 09:30 |
MTDiscord |
<siliconsniffer> Personally, I am fine with leaving that last thing out of the highlights, as there is no visual feedback for the user, and it feels half-baked. |
| 14:13 |
|
mrcheese joined #luanti-dev |
| 15:34 |
|
Awkanimus joined #luanti-dev |
| 16:07 |
sfan5 |
merging #16863, #16853, #16818, #16811 in 15m |
| 16:07 |
ShadowBot |
https://github.com/luanti-org/luanti/issues/16863 -- Fix typo in option desc by AMDmi3 |
| 16:07 |
ShadowBot |
https://github.com/luanti-org/luanti/issues/16853 -- Add translator comments for some ambiguous strings by Wuzzy2 |
| 16:07 |
ShadowBot |
https://github.com/luanti-org/luanti/issues/16818 -- Add 'curl_proxy' setting by j-r |
| 16:07 |
ShadowBot |
https://github.com/luanti-org/luanti/issues/16811 -- Reduce send block loops at large distances by lhofhansl |
| 16:42 |
|
gildarts joined #luanti-dev |
| 17:09 |
rubenwardy |
what new shaders are supported on Android now vs 5.14.0? I see the PR but it doesn't list them |
| 17:09 |
rubenwardy |
other than dynamic shadows |
| 17:13 |
sfan5 |
everything gated behind enable_dynamic_shadows |
| 17:14 |
|
wsor4035 joined #luanti-dev |
| 17:16 |
user333_ |
dynamic shadows look amazing but i keep them off because it isn't worth it for 45 FPS, though i assume a modern phone could handle it |
| 17:20 |
sfan5 |
rubenwardy: this is the full-size screenshot in case you want to use it https://forum.luanti.org/download/file.php?id=31163 |
| 17:30 |
rubenwardy |
enable_water_reflections - i feel like this should be called specular highlights or something as it's not true reflections |
| 17:31 |
rubenwardy |
so it appears to juse be enable_water_reflections and translucent foilage |
| 17:36 |
|
[MatrxMT] joined #luanti-dev |
| 17:41 |
|
[MatrxMT] joined #luanti-dev |
| 19:06 |
MTDiscord |
<luatic> irrlicht never ceases to amaze. apparently quaternion multiplication basically has its operands swapped. |
| 19:06 |
MTDiscord |
<luatic> naturally, this does not seem to be documented anywhere, and is inconsistent with matrices.# |
| 19:06 |
celeron55 |
good enough for quake clones |
| 19:07 |
MTDiscord |
<luatic> i mean really i've seen my fair share of notation, i don't mind. in principle it's completely valid to want to do function composition the other way around. this is also what stacklangs / RPN do. |
| 19:08 |
MTDiscord |
<luatic> but for the love of god, if you do this, do it consistently, and document it well because it is very nonstandard |
| 19:21 |
Krock |
to be fair, Luanti API functions aren't always consistent either. |
| 21:18 |
rubenwardy |
are texture arrays supports on android / gles? |
| 21:18 |
sfan5 |
yes |
| 21:19 |
sfan5 |
requires GLES 3.0, which most people have |
| 21:19 |
rubenwardy |
okay, will the driver be gles3 then? |
| 21:19 |
rubenwardy |
*ogles3 |
| 21:19 |
rubenwardy |
I only see ogles2 in the list |
| 21:20 |
sfan5 |
ogles2 |
| 21:20 |
sfan5 |
...because it supports OpenGL ES 2.0 and later |
| 21:22 |
rubenwardy |
fact check: |
| 21:22 |
rubenwardy |
> Your device needs to support and use OpenGL 3 or GLES 3. You can verify this by checking the About tab of the main menu. On desktop, it should say “opengl3” and not “opengl” on the bottom left. On Android, it should say “ES 3 / ogles2”. |
| 21:23 |
sfan5 |
correct |
| 21:24 |
sfan5 |
very minor point: our 'opengl3' driver requires OpenGL 3.2 |
| 21:24 |
rubenwardy |
will update that in the first sentence |
| 21:25 |
sfan5 |
by the way dynamic shadows require at least OpenGL ES 3.2 |
| 21:26 |
sfan5 |
(could possibly work on ES 3.0 with the right extensions, which makes it hard to say this exactly) |
| 21:30 |
rubenwardy |
does texture arrays require 3.2 as well or just 3.0? |
| 21:33 |
sfan5 |
just 3.0 |
| 21:33 |
user333_ |
i sucessfully ran 5.14 with a machine that could only do OpenGL 3.1 |
| 21:33 |
sfan5 |
OpenGL ES 3.0 *or* OpenGL 3.2 (since that are the base requirements for the 'opengl3' driver) |
| 21:33 |
sfan5 |
to be clear |
| 21:34 |
user333_ |
but it didn't meet either of those requirements- it couldn't even do 3.3 which was required for software like PrusaSlicer and Blender |
| 21:36 |
sfan5 |
we're not talking about the minimum requirements. which haven't changed by the way. |
| 21:44 |
rubenwardy |
anyone noticed the macos builds are named differently |
| 21:44 |
rubenwardy |
anyway Android builds uploaded, doing playstore now |
| 21:44 |
rubenwardy |
this is so much less painful with the script |
| 21:47 |
rubenwardy |
ok in review |
| 21:48 |
rubenwardy |
draft blog post https://github.com/luanti-org/blog/pull/173 |
| 21:54 |
sfan5 |
commented |
| 22:01 |
rubenwardy |
what's the difference between an array texture and a texture array? |
| 22:01 |
user333_ |
the order of two words |
| 22:02 |
sfan5 |
https://wikis.khronos.org/opengl/Array_Texture they're called that |
| 22:02 |
rubenwardy |
okay so is everyone who calls it a texture array wrong |
| 22:02 |
sfan5 |
technically yes |
| 22:03 |
rubenwardy |
ie https://www.youtube.com/watch?v=YTfdBSjitd8 |
| 22:03 |
sfan5 |
i guess it depends on whether you use "texture" to refer to the logical or physical texture (for lack of better words) |
| 22:04 |
sfan5 |
a $whatever_you_call_it is a single logical texture but contains multiple physical textures |
| 22:04 |
sfan5 |
wait I think that's not helpful at all |
| 22:07 |
rubenwardy |
updated |
| 22:08 |
rubenwardy |
a texture array is inside an ArrayTexture |
| 22:08 |
rubenwardy |
luckily we don't need to explain this |
| 22:11 |
rubenwardy |
actually I think auri is using vulkan there |
| 22:35 |
|
wsor4035 joined #luanti-dev |
| 23:01 |
|
user333_alt joined #luanti-dev |
| 23:32 |
|
panwolfram joined #luanti-dev |
