| Time |
Nick |
Message |
| 00:57 |
|
olliy joined #minetest-dev |
| 01:11 |
|
olliy_ joined #minetest-dev |
| 04:17 |
|
cheapie_ joined #minetest-dev |
| 04:25 |
|
proller joined #minetest-dev |
| 06:25 |
|
ssieb joined #minetest-dev |
| 08:00 |
|
ShadowNinja joined #minetest-dev |
| 08:23 |
|
calcul0n joined #minetest-dev |
| 08:26 |
|
hhmhm joined #minetest-dev |
| 11:04 |
|
Fixer joined #minetest-dev |
| 11:31 |
|
calcul0n_ joined #minetest-dev |
| 12:37 |
|
karamel joined #minetest-dev |
| 12:44 |
|
Warr1024 joined #minetest-dev |
| 12:46 |
|
Warr1024 joined #minetest-dev |
| 12:52 |
|
MarwolTuk joined #minetest-dev |
| 13:05 |
|
olliy joined #minetest-dev |
| 13:07 |
|
proller joined #minetest-dev |
| 13:32 |
|
karamel joined #minetest-dev |
| 13:53 |
|
karamel joined #minetest-dev |
| 14:29 |
Jordach |
sfan5, managed to get first person attachments working with full backwards api compat |
| 14:50 |
|
karamel joined #minetest-dev |
| 15:00 |
|
jonadab joined #minetest-dev |
| 15:58 |
MTDiscord |
<appguru> "first-person attachments" |
| 15:58 |
MTDiscord |
<appguru> by that you mean attachments which are visible in first person? |
| 16:03 |
MTDiscord |
<Jordach> yes |
| 16:04 |
Jordach |
appguru: https://gfycat.com/welloffillfatedemperorshrimp |
| 16:05 |
MTDiscord |
<appguru> really noice |
| 16:05 |
Jordach |
set_bone_position and set_animation are used together for this |
| 16:05 |
Jordach |
so fake wields are now acceptable, as are usable arms |
| 16:06 |
|
Lunatrius joined #minetest-dev |
| 17:04 |
|
Warr1024 joined #minetest-dev |
| 17:39 |
|
proller joined #minetest-dev |
| 17:46 |
|
karamel joined #minetest-dev |
| 17:46 |
|
MarwolTuk joined #minetest-dev |
| 17:46 |
|
Taoki joined #minetest-dev |
| 17:46 |
|
Sokomine joined #minetest-dev |
| 17:46 |
|
ghoti joined #minetest-dev |
| 17:46 |
|
TC01 joined #minetest-dev |
| 18:03 |
sfan5 |
merging #10329 and #10185 in 5m |
| 18:03 |
ShadowBot |
https://github.com/minetest/minetest/issues/10329 -- Mark multiply and divide with two vectors as deprecated (Schur product and quotient) by Desour |
| 18:03 |
ShadowBot |
https://github.com/minetest/minetest/issues/10185 -- Change default ambient occlusion gamma to 1.8 by sfan5 |
| 18:06 |
MTDiscord |
<appguru> please no |
| 18:06 |
|
appguru joined #minetest-dev |
| 18:06 |
appguru |
I would like to object to #10329 |
| 18:06 |
ShadowBot |
https://github.com/minetest/minetest/issues/10329 -- Mark multiply and divide with two vectors as deprecated (Schur product and quotient) by Desour |
| 18:07 |
appguru |
That PR suggests just deprecating widely used and useful functionality without even offering a replacement |
| 18:07 |
sfan5 |
in what way? |
| 18:07 |
sfan5 |
where exactly is that widely used? |
| 18:08 |
celeron55 |
this is already discussed in the PR |
| 18:08 |
appguru |
sorcerykid is giving one example, I remember using that quite some times |
| 18:08 |
sfan5 |
the PR no longer depractes + and - like you seem to be assuming |
| 18:08 |
sfan5 |
deprecates* |
| 18:09 |
appguru |
I know, it just deprecates vector.divide(a, b) and multiply with a, b vectors |
| 18:10 |
appguru |
I don't see any good reason to mark them as deprecated, the only one would probably be performance |
| 18:11 |
rubenwardy |
I see an issue when it comes to implementing `*`, but I don't care that much about `multiply` |
| 18:11 |
appguru |
If performance was a main goal, you should focus on deprecating the {x = ..., y = ..., z = ...} format |
| 18:11 |
appguru |
is* |
| 18:11 |
sfan5 |
"not supporting functionality nobody needs" is enough of a reason IMO |
| 18:12 |
appguru |
Give me a second, I'll count my vector.multiply and divide usages in my mods |
| 18:12 |
sfan5 |
(of course this PR doesn't do that, it just paves the way by deprecating it) |
| 18:12 |
|
ssieb joined #minetest-dev |
| 18:12 |
sfan5 |
sure, tell me if you find a schur product or quotient |
| 18:12 |
appguru |
I'll for sure find one |
| 18:14 |
appguru |
Ah yes, here's my first example, something that should be found in moblib: |
| 18:14 |
appguru |
clicker:set_properties{visual_size = vector.divide(clicker:get_properties().visual_size, obj:get_properties().visual_size)} |
| 18:14 |
appguru |
I use this on entities with variable visual size, as attachments are (in this case undesirably) scaled by the visual size as well |
| 18:15 |
appguru |
So I have to set the player visual size = player visual size / entity visual size |
| 18:15 |
appguru |
For that I need the Schur quotient, and to reverse it I need the Schur product |
| 18:16 |
celeron55 |
it's the wrong name for the operation nevertheless |
| 18:17 |
appguru |
Separating the names is what this PR should do |
| 18:17 |
celeron55 |
if you open the wikipedia page for vector multiplication, it doesn't even list schur product |
| 18:17 |
sfan5 |
worldedit code has a schur product too, interesting |
| 18:17 |
sfan5 |
anyway merging those two PRs |
| 18:18 |
appguru |
Only other example I found is from Voxelizer: local transform_vec = vector.divide(mt_space, obj_space) |
| 18:18 |
appguru |
Basically it does some scaling |
| 18:18 |
appguru |
vector.multiply should probably be called vector.scale for two vectors |
| 18:18 |
appguru |
hmmm |
| 18:19 |
appguru |
vector.divide would then be vector.rescale or something |
| 18:19 |
celeron55 |
it could be added as vector.scale if we don't want to use mathematical terms that modders generally don't know |
| 18:19 |
appguru |
schur_product would make it look scarier than it is |
| 18:19 |
appguru |
far scarier* |
| 18:23 |
|
proller joined #minetest-dev |
| 18:26 |
|
GreenXenith joined #minetest-dev |
| 18:36 |
|
fluxflux joined #minetest-dev |
| 18:42 |
|
proller joined #minetest-dev |
| 18:48 |
|
MTDiscord joined #minetest-dev |
| 18:48 |
|
homthack joined #minetest-dev |
| 19:10 |
Jordach |
sfan5, got time to review #10360 today? |
| 19:10 |
ShadowBot |
https://github.com/minetest/minetest/issues/10360 -- Fix/Add First Person Attachments by Jordach |
| 19:11 |
sfan5 |
I scrolled through it earlier and wondered if "forced visible" was the best way to name it, there were also a few minor style issues |
| 19:12 |
Jordach |
it's a weird hack but it just works even with ancient mods |
| 19:15 |
|
absurb joined #minetest-dev |
| 19:15 |
celeron55 |
that's a good point, why isn't it called show_in_first_person |
| 19:16 |
celeron55 |
or visible_ |
| 19:18 |
celeron55 |
elsewhere in the code also |
| 19:28 |
|
proller joined #minetest-dev |
| 19:46 |
|
v-rob joined #minetest-dev |
| 19:57 |
Jordach |
thank god i'm not having to touch unit_sao again |
| 19:58 |
Jordach |
now it's just a slight issue with client side rendering |
| 20:33 |
Jordach |
...is it a hack to mark the local player CAO as front face culling? |
| 20:36 |
|
homthack joined #minetest-dev |
| 20:45 |
pgimeno |
hm, about #10329, Schur product is very often used in GLSL to tint colours. I wonder how often colour components are represented by a vector in Minetest. |
| 20:45 |
ShadowBot |
https://github.com/minetest/minetest/issues/10329 -- Mark multiply and divide with two vectors as deprecated (Schur product and quotient) by Desour |
| 20:47 |
pgimeno |
not only that, but * in GLSL is the Schur product |
| 20:48 |
pgimeno |
you do dot() or cross() if you want some other kind of product |
| 21:01 |
pgimeno |
and in fact, some shaders in Minetest uses Schur product, e.g. client/shaders/nodes_shader/opengl_fragment.glsl lines 60-61, line 207 |
| 21:04 |
Jordach |
sfan5, yep, that fixed it |
| 21:05 |
pgimeno |
to my surprise, even the vertex shader does, see client/shaders/nodes_shader/opengl_vertex.glsl line 58 |
| 21:25 |
|
jonadab joined #minetest-dev |
| 21:26 |
Jordach |
sfan5, fix pushed |
