| Time |
Nick |
Message |
| 00:02 |
|
turtleman joined #luanti-dev |
| 00:20 |
MTDiscord |
<luatic> backwards compat. it's a planned breakage. |
| 01:29 |
|
Juri joined #luanti-dev |
| 01:29 |
|
Niklp2 joined #luanti-dev |
| 01:36 |
|
Niklp2 joined #luanti-dev |
| 01:36 |
|
Juri joined #luanti-dev |
| 03:14 |
|
Juri joined #luanti-dev |
| 03:14 |
|
Niklp2 joined #luanti-dev |
| 04:00 |
|
MTDiscord joined #luanti-dev |
| 05:20 |
|
Juri joined #luanti-dev |
| 05:20 |
|
Niklp2 joined #luanti-dev |
| 09:42 |
MTDiscord |
<the4spaceconstants2181> backwards compat in what way? the meshes? well that can be kept (but still though why is the velocity limit in-code written in 10x) |
| 09:57 |
celeron55 |
you're not going to believe this but luanti's internal coordinate system is such that 1 node = 10 units |
| 09:57 |
celeron55 |
when it comes to physics and rendering |
| 09:57 |
celeron55 |
or, at least it used to be |
| 09:58 |
celeron55 |
i'm not sure how much of it has been changed from that so far, but if you see a factor of 10 anywhere, that's the one and only reason |
| 09:59 |
celeron55 |
and the name for the 10x factor is "BS" |
| 10:00 |
celeron55 |
or, to put it in another way, the dimension of a node is BS |
| 10:00 |
celeron55 |
however, the modding API didn't adapt this convention and simplified it to node = 1 |
| 10:01 |
celeron55 |
adopt* |
| 10:13 |
MTDiscord |
<the4spaceconstants2181> (are you addressing me or luatic) yeah I knew that about the codebase, I'm asking to change that while keeping it the same where there's an interface with mods |
| 11:26 |
[MatrxMT]_ |
<Zughy> The dimension of a node is bullshit? Wow, what a joke of an engine |
| 11:26 |
[MatrxMT]_ |
<Zughy> Well, game platform, whatever |
| 11:26 |
rubenwardy |
this isn't uncommon but other engines do a better job of not leaking it |
| 11:27 |
rubenwardy |
well, *unheard of |
| 11:27 |
rubenwardy |
definitely should have been fixed with 5.0 |
| 13:44 |
|
loggingbot_ joined #luanti-dev |
| 13:44 |
|
Topic for #luanti-dev is now Luanti (fka Minetest) core development and maintenance. Chit-chat goes to #luanti | https://dev.luanti.org/ https://irc.luanti.org/ https://github.com/luanti-org |
| 14:44 |
MTDiscord |
<luatic> apparently going compat -> core means we can directly get OGL 3 on macOS, even without ANGLE |
| 14:46 |
sfan5 |
its not a huge amount of work, too |
| 14:48 |
MTDiscord |
<luatic> indeed. i just slopped a prototype with claude fable, diff came out to 20 files changed, 190 insertions, 268 deletions. |
| 14:55 |
sfan5 |
who is paying for that? :D |
| 14:58 |
MTDiscord |
<luatic> i have a personal claude code pro subscription and presumably as part of some marketing stunt anthropic lets mere pro users try fable a bit (until July 12th for me) so that's what i'm doing |
| 15:28 |
|
fgaz joined #luanti-dev |
| 15:29 |
|
jfred_ joined #luanti-dev |
| 16:31 |
|
crazylad joined #luanti-dev |
| 17:57 |
|
crazylad joined #luanti-dev |
| 18:16 |
|
crazylad joined #luanti-dev |
| 18:33 |
MTDiscord |
<.zenonseth> does your claude usage randomly reset sometimes? Mine's done it 4-5 weeks now, not the same day every week and 3 times (including just now) went back to 0% for the week, but once it went form like 40% to 20% |
| 20:55 |
|
crazylad joined #luanti-dev |
| 21:12 |
MTDiscord |
<.zenonseth> has SSM<->SSCSM communication already been discussed? I can do more of the low hanging fruit stuff, like particles API, but I'm curious if that was already discussed and/or ideas exist, because I can also have a go at it, even better if ideas already exist n any nuances |
| 21:32 |
MTDiscord |
<luatic> Zenon Seth: i think the obvious solution would be extending mod channels (already a thing in CSM) to SSCSM, no? |
| 22:19 |
MTDiscord |
<.zenonseth> I'm actually not that familiar with CSM mods to behind with, looking at the ModChannelMgr, that looks like subscribe/publish type of thing, that anyone can potentially write to - i assume we'd want to limit mods to only talk to their own SSCSMs? That actually makes me wonder, is the plan to allow a mod to have multiple SSCSMs or just 1 per mod? |
| 22:22 |
|
diceLibrarian joined #luanti-dev |
| 22:26 |
MTDiscord |
<luatic> the current architecture is that there will be exactly one SSCSM per mod |
| 22:28 |
MTDiscord |
<luatic> you do want multiple channels, in particular because you might want to map them to actual networking channels in the future, and then you don't want independent channels head-of-line blocking each other. |
| 22:28 |
MTDiscord |
<luatic> so then you just end up with an API that lets you publish / subscribe to named channels and that's basically mod channels |
| 22:29 |
MTDiscord |
<luatic> i'd recommend looking at the (brief) API documentation in [client_]lua_api.md to understand how they work from the modder's perspective rather than the verbose C++ implementation |
| 22:29 |
MTDiscord |
<luatic> once you have named channels, it is no trouble to set up a named channel per-mod, if that's what you want |
| 22:30 |
MTDiscord |
<.zenonseth> ok, I'll have a look at the docs - i'm assuming these channels aren't going to be directly modifyable by the mods themselves, otherwise they could send commands to other mods' sscsm - |
| 22:31 |
MTDiscord |
<luatic> they could. i don't really see that as a big problem (yet). |
| 22:31 |
MTDiscord |
<luatic> "private" channels could always be added later, if SSCSM ever gets some sort of "trust" system. |
| 22:32 |
|
panwolfram joined #luanti-dev |
| 22:34 |
MTDiscord |
<.zenonseth> yeah i guess there's already global shared namespace so mods can modify other ssms - i guess then the only thing here is the channel collision, if users can freely name them, maybe just auto-namespace mod channels with modname: prefix unless a mod specifically requests to send join othermod:something |
| 22:36 |
|
mrcheese joined #luanti-dev |
| 22:37 |
MTDiscord |
<.zenonseth> i think maybe this is also best left for later when sscsms can actually be sent by mods then, no need for me to rush, but good info to know! |
