| Time |
Nick |
Message |
| 00:00 |
MTDiscord |
<the4spaceconstants2181> no markdown in general afaik |
| 00:00 |
MTDiscord |
<the4spaceconstants2181> that's client-dependent |
| 00:01 |
MTDiscord |
<wsor4035> yes, but i kinda figured it would send ~~ at the start and end at least |
| 00:02 |
MTDiscord |
<the4spaceconstants2181> is it treating each one as ~.....~? |
| 00:18 |
user333_ |
i believe you can do strikethroughs, among other types of formatting |
| 04:00 |
|
MTDiscord joined #luanti-dev |
| 04:29 |
|
YuGiOhJCJ joined #luanti-dev |
| 04:50 |
|
YuGiOhJCJ joined #luanti-dev |
| 06:23 |
|
MTDiscord joined #luanti-dev |
| 06:30 |
|
YuGiOhJCJ joined #luanti-dev |
| 07:43 |
|
YuGiOhJCJ joined #luanti-dev |
| 08:08 |
sfan5 |
yeah that works here |
| 08:26 |
|
YuGiOhJCJ joined #luanti-dev |
| 09:12 |
|
YuGiOhJCJ joined #luanti-dev |
| 09:43 |
|
ivanbu joined #luanti-dev |
| 12:31 |
|
turtleman joined #luanti-dev |
| 12:50 |
|
turtleman joined #luanti-dev |
| 14:35 |
MTDiscord |
<wsor4035> for #10371, to prevent modders from locking in on one specific db, what if luanti where to provide a drizzle "orm" like api? i say orm in "" since if you know sql it takes all of 3 seconds to learn drizzle. obviously wouldn't have any type safety, etc, but it would be an api that could be wrapped over any db. sample query: const result = await db.select().from(users).where(eq(users.id, 1)); docs: https://orm.drizzle.team/ |
| 14:35 |
ShadowBot |
https://github.com/luanti-org/luanti/issues/10371 -- Lua API: Access to any DBMS |
| 16:00 |
|
pmp-p joined #luanti-dev |
| 16:14 |
|
ivanbu joined #luanti-dev |
| 17:07 |
sfan5 |
well who is going to invest time writing it |
| 17:14 |
MTDiscord |
<wsor4035> The api itself? It's very much defined on both end spec, so one could probably trivially use an llm to create it and the check/test the output |
| 17:15 |
sfan5 |
well uh |
| 17:16 |
sfan5 |
I can't say I disagree with that approach in principle, but we are just about to define a policy that would ban this |
| 17:17 |
MTDiscord |
<wsor4035> Actually no. With the changes its a might be closed |
| 17:17 |
MTDiscord |
<wsor4035> Average luanti stupidity is what that pr is tho |
| 17:18 |
Sheriff_U3 |
It's wasting core dev time more then a AI slop pr at this point I think. |
| 17:19 |
MTDiscord |
<wsor4035> > Generating substantial amounts of code with AI is strongly discouraged and > may result in immediate closure of the PR |
| 17:19 |
MTDiscord |
<wsor4035> note the may |
| 17:19 |
MTDiscord |
<wsor4035> yes |
| 17:20 |
MTDiscord |
<wsor4035> also the only actual closed pr is someone trying to slop a bug bounty |
| 17:20 |
MTDiscord |
<wsor4035> everything else has been a witch hunt |
| 17:23 |
MTDiscord |
<wsor4035> anyways, back to sql things, i said two defined specs, technically its three, since you would need something to control which sql your mapping to under the hood. this probably is slightly less defined, as iirc (citation needed) luantis current support for multiple sql backends is just a glorifed kv wrapper |
| 17:30 |
MTDiscord |
<wsor4035> the one thing to note is luanti currently has everything under core/minetest global.func. for something like this it might make sense to have it as its own thing that could be required, which would mean https://github.com/luanti-org/luanti/pull/16036 would have to be merged first. course doing this would mean more classic luanti politics/bikeshedding tho |
| 17:33 |
|
crazylad joined #luanti-dev |
| 17:56 |
|
Desour joined #luanti-dev |
| 19:00 |
Krock |
For some reason, anisotropic_filter changes the appearance of the entire node regardless on the use_alpha_texture field. https://gcdnb.pbrd.co/images/FSBlHEytJ0aA.gif |
| 19:01 |
sfan5 |
cool |
| 19:04 |
Desour |
wow, it even rotates the node |
| 19:04 |
Krock |
crazy indeed |
| 19:07 |
Krock |
also when "mip_map = true" |
| 19:20 |
Krock |
HUH it fills up the transparent pixels with whatever color is nearby |
| 19:22 |
Krock |
sfan5: didn't you once post an image where you used the default_apple.png and filled its colors outwards? I remember seeing something like that .... it was a while ago. |
| 19:22 |
sfan5 |
yes |
| 19:22 |
sfan5 |
that's the imageCleanTransparent algorithm |
| 19:23 |
Krock |
> if (m_setting_mipmap || m_setting_bilinear_filter || |
| 19:23 |
Krock |
m_setting_trilinear_filter || m_setting_anisotropic_filter) { |
| 19:23 |
Krock |
now that explains the different outcomes! Good to know, thanks. |
| 19:24 |
sfan5 |
we could simply state in the documentation that modders should not rely on alpha turning to any specific color (black) when using a transparent texture with a non-transparent node |
| 19:24 |
Krock |
(for context: I was somewhat confused because I couldn't find any node which looked like this https://forum.luanti.org/viewtopic.php?p=451971#p451971 ) |
| 19:28 |
Krock |
is "blend" perhaps not supported on mesh nodes? |
| 19:30 |
sfan5 |
should be |
| 19:31 |
Krock |
sorry. typo in the nodedef field. |
| 19:48 |
Krock |
will merge #16846, #17336, #17337 in 15 minutes |
| 19:48 |
ShadowBot |
https://github.com/luanti-org/luanti/issues/16846 -- Formspec: Make inventory slots stay hovered on formspec update by CrazyladMT |
| 19:48 |
ShadowBot |
https://github.com/luanti-org/luanti/issues/17336 -- Mainmenu: Progress bar show more media download details by cx384 |
| 19:48 |
ShadowBot |
https://github.com/luanti-org/luanti/issues/17337 -- Main menu: Mod sorting and display improvements by SmallJoker |
| 20:05 |
Krock |
- done - |
| 20:19 |
Sokomine |
tab name completition now adds a blank after the name. that's impractical. configurable? |
| 20:23 |
Krock |
I suppose it's impractical when you'd like to add a punctuation character afterwards |
| 20:25 |
Krock |
however I suspect you'd also be uncertain whether the autocompletion is actually done. If you'd like to have a setting for that, it could be implemented. Though I am not sure whether it will be more practical that way. |
| 20:36 |
Sokomine |
i use it often to greet all players who are online. that does require a comma after each name. i also type the space automaticly after pressing tab so i end up with two |
| 20:38 |
Krock |
noted. will write a PR tomorrow |
| 20:50 |
Sokomine |
thanks! |
| 21:15 |
sfan5 |
OBJ mesh parsing seems to be relatively slow https://x0.at/o5gL.png |
| 21:15 |
sfan5 |
or alternative explanation: most meshes in this game session are obj |
| 22:24 |
sfan5 |
is there any reason why we haven't defined std::min/max (or similar) for our vector types |
| 22:24 |
sfan5 |
or am I missing these methods? |
| 22:35 |
|
panwolfram joined #luanti-dev |
| 22:47 |
|
crazylad joined #luanti-dev |
| 23:13 |
MTDiscord |
<josiah_wi> It would be std::min/max_element for a container, right? |
| 23:14 |
MTDiscord |
<josiah_wi> Oh, but it would be std::min/max for a mathematical vector type. Right. |
| 23:14 |
Desour |
there is no obvious min/max of 2 vectors |
| 23:16 |
sfan5 |
well there's (a < b ? a : b) |
| 23:16 |
sfan5 |
but it turned out that I actually wanted componentwise_min |
| 23:17 |
Desour |
there's also no obvious < on vectors ;) |
| 23:17 |
Desour |
is defining std::min even allowed? |
| 23:17 |
sfan5 |
no |
| 23:18 |
sfan5 |
but we can have our own helper of course |
| 23:19 |
Desour |
ah, that's how you meant it. I don't see a reason against adding a componentwise_min for vectors |
| 23:20 |
sfan5 |
(that exists already) |
| 23:21 |
Desour |
(but you could also add something more general: an `auto apply<F>(vector v1, vector v2, F func)`, so it would be `v3 = v3f::apply(v1, v2, std::min)`) |
| 23:22 |
Desour |
s/apply/combine/ |
| 23:30 |
MTDiscord |
<luatic> engine vectors should have all the things lua vectors have, yeah |
| 23:31 |
MTDiscord |
<luatic> btw sfan5 for your earlier question, static models that have set_animation() or play_animation() called on them will now produce warnings, yes |
| 23:31 |
MTDiscord |
<luatic> i considered this a feature at the time but i'm fine if it's disabled for static models |
