| Time |
Nick |
Message |
| 00:07 |
|
duds` joined #minetest |
| 00:07 |
|
Peter_Lankton joined #minetest |
| 00:09 |
Peter_Lankton |
so i am still kind of new to lua and minetest modding. is there a way to override the objectref:set_pos() function? i want to make it so if any mod teleports a player, it first does something then runs the original function i.e. teleport the player. |
| 00:37 |
|
Guest87 joined #minetest |
| 00:40 |
|
duds` joined #minetest |
| 00:54 |
|
Alnotz joined #minetest |
| 01:09 |
|
Lesha_Vel joined #minetest |
| 01:14 |
|
duds` joined #minetest |
| 01:17 |
Kimapr |
Peter_Lankton: Yes. you need to acquire one (1) ObjectRef, then modify its methodtable (pass the obj to `getmetatable` to get it) |
| 01:17 |
|
smk joined #minetest |
| 01:29 |
|
Sven_vB joined #minetest |
| 01:36 |
Peter_Lankton |
Kimapr it tells me its a nil value when i try "ObjectRefmetatable = ObjectRef:getmetatable()" |
| 01:39 |
Peter_Lankton |
oh wait no i think i figured out what i did wrong lol |
| 01:42 |
Peter_Lankton |
i had thought the metatable func was part of the object thingy. anyway, with that done, i now have the best anti-noclip mod ever. |
| 01:44 |
germ |
So that framinetest server ? |
| 01:44 |
germ |
Peeps get educated, much ? |
| 01:44 |
Peter_Lankton |
its still buggy and sometimes makes the player get stuck. however i think i can fix that. its already the best regardless. |
| 01:45 |
Peter_Lankton |
i have no idea what you are talking about. i dont play other peoples servers. no need to be a dick about it |
| 01:48 |
germ |
Lol no! |
| 01:49 |
Peter_Lankton |
ok then link me the source repo of this better anti noclip mod. |
| 01:49 |
|
duds` joined #minetest |
| 01:49 |
germ |
I wasn't even reading, it's a cold prompt: I am trying to get the kid of a friend `educated' (whatever this means) |
| 01:50 |
Peter_Lankton |
oh were u replying to an old post? |
| 01:50 |
|
YuGiOhJCJ joined #minetest |
| 01:51 |
germ |
Not even ! I just had a tried at that Framinetest server.. I'm curious where I could take it. Idk if there is a lot more than what I've seen so far |
| 01:52 |
Peter_Lankton |
oh someone named Peeps im guessing. i guess i was not aware because peeps also means people. or do u mean u acidentally pasted that here instead of the game chat? |
| 01:53 |
germ |
I meant people |
| 01:53 |
Peter_Lankton |
i thought u talking about the framinetest thing was responding to me about my noclip mod. telling me they have a better one. |
| 01:54 |
germ |
no not at all.. You have a problem with your Minetest, comrade ? |
| 01:55 |
Peter_Lankton |
oh ok. sry for the misunderstanding lol. no im good. i mean i was having problems but i figured out a solution. which was why i came here in the first place. |
| 01:56 |
germ |
oh that's neat .. Yeah I should read more before just starting talking.. I am so alone lol |
| 01:56 |
Peter_Lankton |
heh. me too |
| 02:03 |
Peter_Lankton |
so its probably not the first. but my anti noclip mod uses the raycast feature. i tried looking for other anti noclip mods but they kinda sucked. so i am making my own. right now any node that has a collision that is different from the nodebox has issues as the raycast seems to use the nodebox. so maybe it just will check and ignore nodes where the nodebox and collisionbox dont match. it also has to check whether the |
| 02:03 |
Peter_Lankton |
node is collidable anyway. |
| 02:06 |
|
Sven_vB joined #minetest |
| 02:07 |
Peter_Lankton |
heh im so lonely now that im discussing my mod development in the irc. which i dont usually do. anyway since its just gonna distract me and i got what i came here for, i will be going now. byeeeeeeeee! |
| 02:07 |
|
Peter_Lankton left #minetest |
| 02:09 |
muurkha |
byee |
| 02:16 |
|
Trifton joined #minetest |
| 02:26 |
|
duds` joined #minetest |
| 03:03 |
|
duds` joined #minetest |
| 03:43 |
|
duds` joined #minetest |
| 03:47 |
|
peterz joined #minetest |
| 04:00 |
|
MTDiscord joined #minetest |
| 04:12 |
|
peterz joined #minetest |
| 04:16 |
|
peterz joined #minetest |
| 04:19 |
|
duds` joined #minetest |
| 04:49 |
|
Lesha_Vel joined #minetest |
| 04:53 |
|
duds` joined #minetest |
| 05:43 |
|
calcul0n joined #minetest |
| 06:05 |
|
appguru joined #minetest |
| 06:14 |
|
duds` joined #minetest |
| 06:38 |
|
TomTom_ joined #minetest |
| 07:32 |
|
appguru joined #minetest |
| 07:55 |
|
duds` joined #minetest |
| 08:19 |
|
mrkubax10 joined #minetest |
| 08:21 |
|
Thelie joined #minetest |
| 08:29 |
|
duds` joined #minetest |
| 08:44 |
|
imi joined #minetest |
| 08:44 |
imi |
hi, am I allowed to double reference a stack in two inventory lists or am I supposed to make a deep copy? |
| 09:03 |
|
duds` joined #minetest |
| 09:09 |
|
Thelie joined #minetest |
| 09:32 |
MTDiscord |
<luatic> Peter_Lankton: Overriding :set_pos doesn't help with detecting noclip at all, unless the "noclip" is induced by a bug in one of your mods calling :set_pos on the player. Player movement does not call :set_pos. |
| 09:32 |
MTDiscord |
<luatic> ah nevermind that was unrelated apparently |
| 09:33 |
MTDiscord |
<luatic> btw raycast uses the selectionbox, not the collisionbox, so yeah, that doesn't work |
| 09:38 |
|
duds` joined #minetest |
| 09:46 |
|
appguru joined #minetest |
| 10:00 |
|
sys4_ joined #minetest |
| 10:39 |
|
mrkubax10 joined #minetest |
| 10:44 |
|
proller joined #minetest |
| 10:53 |
|
mrkubax10 joined #minetest |
| 11:16 |
|
pomup joined #minetest |
| 11:40 |
|
appguru joined #minetest |
| 12:06 |
|
Alnotz joined #minetest |
| 12:43 |
sfan5 |
imi: that's allowed, the api functions just copy the data out of the ItemStacks and don't persist the objects themselves |
| 12:45 |
imi |
ok thanks |
| 13:27 |
|
sparky4 joined #minetest |
| 14:46 |
|
definitelya joined #minetest |
| 15:00 |
|
An[m]1 left #minetest |
| 15:28 |
|
Thelie joined #minetest |
| 15:30 |
|
proller joined #minetest |
| 15:40 |
|
debiankaios joined #minetest |
| 15:58 |
|
mrkubax10 joined #minetest |
| 17:04 |
|
Peter_Lankton joined #minetest |
| 17:05 |
|
meldrian_ joined #minetest |
| 17:05 |
|
potatoxel[m]1 joined #minetest |
| 17:05 |
|
cation_ joined #minetest |
| 17:05 |
|
Thermoriax joined #minetest |
| 17:06 |
|
Thermoriax joined #minetest |
| 17:06 |
|
A_Dragon joined #minetest |
| 17:06 |
|
Talkless joined #minetest |
| 17:06 |
|
vintprox[m] joined #minetest |
| 17:08 |
|
pounce joined #minetest |
| 17:09 |
|
lebruhgamer[m] joined #minetest |
| 17:10 |
|
Fleckenstein[m] joined #minetest |
| 17:10 |
|
tel1 joined #minetest |
| 17:17 |
Warr1024 |
Kimapr and Peter_Lankton: if you want an example of modifying what set_pos does to a player, https://gitlab.com/sztest/nodecore/-/blob/master/mods/nc_player_pickup/teleport.lua |
| 17:18 |
Warr1024 |
The process is pretty simple: I just "juggle" a deferred function every tick until there's at least one player online, then grab the metadata off that player and install the patch. The patch itself just wraps around the old set_pos function and adds the logic I want. |
| 18:33 |
|
illwieckz joined #minetest |
| 18:44 |
|
___nick___ joined #minetest |
| 18:55 |
|
meldrian joined #minetest |
| 19:30 |
|
___nick___ joined #minetest |
| 19:33 |
|
___nick___ joined #minetest |
| 20:12 |
|
Trifton joined #minetest |
| 20:13 |
Peter_Lankton |
@luatic "raycast uses the selectionbox" ok thanks for making that clear to me. i was not sure if it used the selection box or nodebox specifically. i think i will just compromise and make it ignore those nodes. that does weaken the anti noclip a bit but i dont expect houses to be made with those nodes as the walls lol. and Warr1024 thx for the example. i think i found something that works for me. mine will wait until |
| 20:13 |
Peter_Lankton |
the first player joins using the register_on_joinplayer then modify the metatable. just found out i also have to override move_to lol. |
| 20:14 |
MTDiscord |
<luatic> Peter_Lankton: raycasts are the wrong tool anyways |
| 20:14 |
MTDiscord |
<luatic> you probably want box-box collisions, which are pretty simple to implement |
| 20:15 |
Peter_Lankton |
yeah that would be better. |
| 20:15 |
MTDiscord |
<luatic> you likely want something like an adaption of https://github.com/appgurueu/disable_build_where_they_stand, where instead of disabling placement of to-be-placed nodes, a globalstep periodically checks for collisions between players and existing nodes, and deals with players who collide with boxes |
| 20:15 |
MTDiscord |
<luatic> I had considered implementing this but ultimately considered it pointless |
| 20:15 |
MTDiscord |
<luatic> normal players can get stuck in nodes just as well if they are able to dig or place them locally |
| 20:16 |
MTDiscord |
<luatic> so checking for collision unfortunately isn't reliable and distinguishing noclipping from "regular" glitching becomes pretty much impossible |
| 20:17 |
MTDiscord |
<luatic> anti noclip + only considering nodes not diggable/placable by the players + dbwts may work reasonably well though |
| 20:18 |
MTDiscord |
<luatic> TL;DR: if you somehow manage to legitimately enter a block, Minetest allows you to stay and move within it, s.t. noclipping is possible even for non-cheat clients |
| 20:19 |
Warr1024 |
Peter_Lankton: the reason why I didn't use on _joinplayer was because I wanted my hook to (1) automatically expire once I had made the modification to the metatable once, and (2) leave absolutely no footprint once it had been installed. |
| 20:20 |
|
proller joined #minetest |
| 20:20 |
MTDiscord |
<luatic> I don't like that minetest.after loop |
| 20:21 |
MTDiscord |
<luatic> It fails to ensure that the hax are applied as soon as possible |
| 20:21 |
Peter_Lankton |
so i guess i would need some way to cast the shape of the players collision box instead of just a ray cast. i did figure out how to make players not get stuck in nodes they place though. and it kind of works for things like doors too. i just make it ignore the raycast if the above and below are the same pos. |
| 20:21 |
MTDiscord |
<Warr1024> NodeCore does the "if you get stuck inside a solid then you get teleported out" thing too, but it's very limited. It's entirely disabled if you have the noclip priv (it can't check and doesn't care whether its actually "on" or not) and only works on some materials based on full-node collision, not collisionbox (nodes with complex collisionboxes it basically just ignores) and it only works if you stand still for a second (i.e. you have |
| 20:21 |
MTDiscord |
the option to struggle and it will wait for you to be done first). |
| 20:22 |
MTDiscord |
<luatic> Peter_Lankton: "i did figure out how to make players not get stuck in nodes they place though" - that's exactly what dwbts does. And pretty much perfectly. |
| 20:22 |
MTDiscord |
<luatic> (dwbts is short for disable_build_where_they_stand) |
| 20:22 |
MTDiscord |
<Warr1024> I have only one mechanic that avoids using selectionboxes and that one also just uses full nodes (plus, it actually does a custom "old fashioned" raycast with coarse stepping instead, because the ability to "quantum tunnel" is actually desired in that case). |
| 20:22 |
MTDiscord |
<Warr1024> (dwbts is short for disable_where_build_they_stand) |
| 20:23 |
MTDiscord |
<luatic> oof |
| 20:23 |
MTDiscord |
<luatic> double oof, I replied to the wrong message (at least IRC peeps can't see this) |
| 20:23 |
MTDiscord |
<luatic> anyways you don't need coarse stepping for that |
| 20:23 |
MTDiscord |
<Warr1024> I don't need coarse stepping, but it makes it work nice and cheatp |
| 20:24 |
MTDiscord |
<luatic> you can always get the next node by intersecting with the next 3 axis parallel grid planes |
| 20:24 |
MTDiscord |
<luatic> that way you won't miss nodes due to a too coarse stepsize |
| 20:24 |
Peter_Lankton |
dwbts sounds helpful. it is kinda annoying how i can sometimes place blocks inside myself even when noclip mode is off. when its off its inconsistent. |
| 20:24 |
MTDiscord |
<Warr1024> yeah, just adding a vector to another vector is quite a lot less expensive than computing a handful of plane collisions 😄. If I were trying to solve a different problem, maybe ... but then considering the costs of reimplementing raycasts, I'd really try to just pick a different problem entirely. |
| 20:25 |
MTDiscord |
<luatic> it is very manageable, I wrote the code but didn't test it |
| 20:25 |
MTDiscord |
<luatic> it's like 30 lines? |
| 20:25 |
MTDiscord |
<luatic> these planes are all axis parallel, so computing the intersection points is pretty trivial |
| 20:26 |
MTDiscord |
<Warr1024> In my case, I was simulating radiation, and the ability to tunnel through thin films, as well as the ability for non-full-node things to be able to absorb stuff that would have passed through a "hole", were both helpful for both realism and gameplay. |
| 20:26 |
MTDiscord |
<luatic> Peter_Lankton: IIRC the issue is/was that the engine itself uses a dirty clientside hack for enable_build_where_you_stand |
| 20:26 |
MTDiscord |
<luatic> or disable_... |
| 20:26 |
Peter_Lankton |
does it take into account if the player's collision is not default? because in my server, its possible for players to change size and shape. which i know from experience made the mesecons pressure plates look funny. |
| 20:26 |
MTDiscord |
<luatic> (whatever it was called) |
| 20:26 |
MTDiscord |
<Warr1024> It's nice when you can get free realism by just making the bugs in your code match up with the bugs in reality. |
| 20:26 |
MTDiscord |
<luatic> Peter_Lankton: Of course it does. My code is perfect. |
| 20:27 |
Peter_Lankton |
well i will give it a shot |
| 20:28 |
MTDiscord |
<luatic> I don't know why I would be going through 150 lines of computing node collision boxes just to then shittily approximate the player box which Minetest gives me for free 😆 |
| 20:28 |
MTDiscord |
<Warr1024> Using luatic's code ... pros: does what it says pretty well, and even handles a ton of corner cases you probably hadn't thought of. cons: now your project looks like it's been written by a compulsive overengineerer. |
| 20:28 |
MTDiscord |
<luatic> remember me I still have that weight balanced tree in modlib |
| 20:28 |
MTDiscord |
<luatic> reminds me* |
| 20:28 |
MTDiscord |
<Warr1024> hmm, the idea of weight-balanced anything in modlib sounds pretty ironic... |
| 20:29 |
MTDiscord |
<luatic> I don't know if it's actually weight balanced though, I never tested that |
| 20:29 |
MTDiscord |
<luatic> I just tested that it works™️ |
| 20:29 |
MTDiscord |
<luatic> I never checked the weight balance invariant 💀 |
| 20:29 |
MTDiscord |
<luatic> whatever, B-Trees are superior anyways |
| 20:29 |
MTDiscord |
<luatic> I can match the performance of Python's sorted containers lib using them |
| 20:29 |
MTDiscord |
<luatic> (in pure Lua, that is) |
| 20:34 |
Peter_Lankton |
cool it works! though i could not help but notice there is no license on the disable_build_where_they_stand repo. |
| 20:36 |
|
Trifton joined #minetest |
| 20:37 |
MTDiscord |
<Warr1024> ah, upgrade to triple-oof then. |
| 20:45 |
Peter_Lankton |
hmm. also seems to make scaffolding impossible. dunno if thats the right term. but that thing where you hold shift on the side of a edge and place a node under you to build a bridge. cant do that now with the dwbts mod. |
| 20:56 |
|
appguru joined #minetest |
| 21:28 |
MinetestBot |
[git] rubenwardy -> minetest/minetest: Fix shadows dropdown and clean up shader settings (#13481) bc4fc6d https://github.com/minetest/minetest/commit/bc4fc6d6483d37de7703374ce1ade213ff354ec7 (2023-05-03T21:28:02Z) |
| 22:11 |
|
Peter_Lankton left #minetest |
| 22:18 |
|
illwieckz joined #minetest |
| 22:37 |
|
panwolfram joined #minetest |
| 23:51 |
|
Lesha_Vel joined #minetest |
| 23:51 |
|
Charbomber joined #minetest |
