Minetest logo

IRC log for #minetest-dev, 2012-12-16

| Channels | #minetest-dev index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
02:33 parabyte joined #minetest-dev
03:12 VanessaE 22:12:31: ACTION[ServerThread]: Player jojoa1997 tried to access [node under=-50,7,604 above=-50,8,604] from too far: d=1600.43, max_d=140. ignoring.
03:12 VanessaE I'm seeing a LOT of this in my server logs lately, and users complain that they can't place/dig nodes when this is happening.
03:31 VanessaE celeron55: if you're around, please merge those mese changes.  Just got the first new recipe that needs them.
03:32 VanessaE (namely, crafting a UFO that a person can sit in and fly around)
03:32 VanessaE seems somehow fitting given how you view Mese. :-)
05:31 hmmmm joined #minetest-dev
05:53 kotolegokot joined #minetest-dev
06:30 sfan5[iPod] joined #minetest-dev
06:35 Calinou joined #minetest-dev
09:00 rsiska joined #minetest-dev
10:17 PilzAdam joined #minetest-dev
10:20 jin_xi joined #minetest-dev
10:24 rsiska joined #minetest-dev
11:11 darkrose_ joined #minetest-dev
14:51 SpeedProg joined #minetest-dev
15:28 SpeedProg1 joined #minetest-dev
15:55 hmmmm joined #minetest-dev
16:03 serengeor_ joined #minetest-dev
16:03 serengeor joined #minetest-dev
16:04 serengeor joined #minetest-dev
16:36 doserj joined #minetest-dev
17:04 RealBadAngel joined #minetest-dev
17:04 RealBadAngel hi all
17:05 RealBadAngel ive made somethin cool i think, heres the vid on it: http://www.youtube.com/watch?v=GocjjOLuogk
17:16 hmmmm those are great
17:16 PilzAdam yes, they are
17:17 hmmmm i want to start getting my changes into minetest... so right now i'm nearly at the point where my own mapgen is of alpha quality
17:18 hmmmm what i want to happen is that I make a new file, mapgen_legacy.cpp, put all the old stuff in there, turn it into a class, and then replace all the perlin_noise2d() calls with array accesses to the result of my fast perlin noise functions
17:19 hmmmm it'll remain the default mapgen, but you'd be able to switch to the new one if you specify a configuration option like new_mapgen=true or something
17:20 hmmmm if the new mapgen metadata is present in the map_meta.txt or whatever, it'll use the new mapgen regardless of the settings and vice versa
17:20 hmmmm the reason why i want to do this soon is because the branch i'm working with is forked from 0.4.3 and there's no telling how much more difficult it will get to integrate the changes with the master branch as time goes on
17:21 Calinou joined #minetest-dev
17:21 RealBadAngel hi hmmm
17:21 RealBadAngel im makin now gen in c++
17:22 RealBadAngel LUA is far to slow for such stuff
17:22 RealBadAngel im doing matrix4 calculations in lua
17:23 hmmmm yeah, that's what i feared.... you see, "block decorations" such as trees are something you'd obviously want in lua for modifiability but lua seems to be too slow for this sort of stuff
17:24 RealBadAngel it will be customizable by lua
17:24 RealBadAngel gen procedure is the same for all the trees
17:24 hmmmm my master plan was to move the trees out into lua as well, and have a separate mapgen lua interpreter so it'd be able to run in its own thread to not impact the performance of the rest of it
17:25 RealBadAngel just the arguments passed to it are different
17:25 hmmmm oh yeah, you see, that's not going to do it in the long run.... that assumes it's just going to be a tree
17:25 RealBadAngel no
17:25 hmmmm what then....?
17:26 RealBadAngel its L-systems, it can grow any plant
17:26 hmmmm yeah, plant though
17:26 hmmmm i am talking about decorations in general
17:26 hmmmm for example in a nether biome you're not going to want plants to grow
17:26 RealBadAngel so no calls to treegen just
17:27 hmmmm hmm
17:27 RealBadAngel each biome can have defined chances for somethin to spawn
17:27 hmmmm well no, okay, you've got a point
17:27 RealBadAngel so trees = 0
17:28 hmmmm how large is your treegen code
17:28 RealBadAngel wait i will upload most recent
17:28 jin_xi joined #minetest-dev
17:29 RealBadAngel lemme clean it a bit, so main functions will be better visible
17:36 RealBadAngel hmmmm, http://realbadangel.pl/treegen.zip
17:36 RealBadAngel its big because of 128x textures
17:37 RealBadAngel main code is just generate_lsystem_tree = function (pos,initial_axiom,rules_a,rules_b,iterations,angle,bark,leaves)
17:38 RealBadAngel calls with different parameters are in spawners.lua
17:39 RealBadAngel matrix and matrix4 are just libs for calculations on matrices (not needed when inside engine - i rewrote matrix4 stuff based on irrlicht code)
17:40 hmmmm it's all lua ~_~
17:40 RealBadAngel by now
17:40 hmmmm so treegen/matrix is just some math routines
17:40 RealBadAngel i just started to write c++ version
17:41 RealBadAngel yeah
17:41 RealBadAngel thats a lib
17:41 hmmmm did you write it
17:41 RealBadAngel i ported it for sake of treegen
17:41 hmmmm ah
17:41 RealBadAngel its a MIT licensed stuff
17:41 RealBadAngel pretty usable btw
17:42 hmmmm alright so what i'm seeing is a bunch of matrix manipulation stuff along with complex variable functions
17:42 RealBadAngel it will be easier in c++
17:42 hmmmm of course
17:43 hmmmm just be sure to only write the matrix/complex number routines that are being used
17:43 RealBadAngel complex are not needed
17:43 RealBadAngel its just a part of matrix lib
17:44 RealBadAngel im using just in fact 4x4 matrix4f matrices
17:44 RealBadAngel vectors3f, and angles
17:44 hmmmm what routines are needed by treegen
17:44 RealBadAngel and stack for matrices
17:44 hmmmm in the matrix lib
17:44 RealBadAngel add, mul, new
17:44 hmmmm that's all
17:45 hmmmm just code that by han
17:45 hmmmm hand*
17:45 RealBadAngel and those 2 in matrix4 lua
17:45 RealBadAngel setrotation and transpose
17:45 hmmmm yea, no need for a big library just for that
17:45 hmmmm all of which are easy as hell
17:45 RealBadAngel in c++ i dont need them at all
17:45 hmmmm right
17:45 hmmmm okay
17:45 RealBadAngel all is already here
17:45 hmmmm that's good
17:46 RealBadAngel im not sure just bout stacks
17:46 RealBadAngel i need a stack to push/pop matrices
17:46 hmmmm 'stacks'?
17:46 hmmmm you mean the lua stack?
17:46 RealBadAngel no
17:46 RealBadAngel regular stack
17:46 hmmmm oh
17:46 RealBadAngel LIFO
17:46 hmmmm what's wrong with std::stack
17:46 hmmmm http://www.cplusplus.com/reference/stack/stack/
17:47 RealBadAngel i said i wasnt sure just
17:47 RealBadAngel can they stack matrices as well?
17:48 hmmmm sure
17:48 RealBadAngel so no problemo
17:48 hmmmm clearly
17:48 RealBadAngel i got all needed
17:48 RealBadAngel i could go from stark to c++ version
17:49 RealBadAngel but thanks to lua i fully understood how that damn thing works :)
17:49 RealBadAngel including matrix4
17:50 RealBadAngel i havent implemented in LUA all turtle command for lsystems
17:50 RealBadAngel i will add them to c++ version
17:51 RealBadAngel hmmm, btw have you tried already walkin under those trees?
17:51 RealBadAngel i mean in game
17:52 thexyz btw, what's preferable in minetest, STL or Irrlicht containers?
17:52 RealBadAngel idk
17:52 hmmmm STL
17:52 PilzAdam thexyz, c55 dont want to depend too much on Irrlicht
17:52 thexyz hmmmm: why is irrlicht's core::list used everywhere then?
17:52 hmmmm no idea
17:53 RealBadAngel hehehe
17:53 RealBadAngel matrix4 is also irrlicht core
17:53 hmmmm Irrlicht wanted to reimplement all its own containers so it doesn't have a dependency on STL
17:53 PilzAdam dev wiki says dont use it as much as its used in the existing code
17:53 hmmmm but that makes an even bigger dependence on irrlicht
17:53 thexyz hmmmm: yeah, i've that read on theirs blog
17:53 hmmmm and their code has been shown to be shotty
17:53 hmmmm shoddy rather
17:54 hmmmm really it'd be nice if one day someone sat down and switched all the irrlicht containers to STL containers that don't need to be irrlicht containers
17:54 hmmmm but that's boring
17:59 RealBadAngel stupd question, where shall i put lsystem_treegen routines? :)
18:00 thexyz well, if celeron55 approves it, i can spend some time converting irrlicht containers to stl ones, there seems to be not much difference
18:01 hmmmm RealBadAngel, i'd say mapgen.cpp
18:01 hmmmm for now anyway
18:01 RealBadAngel oke
18:06 RealBadAngel btw, epic shot: http://realbadangel.pl/red_cliffs.png
18:07 PilzAdam why not disable the HUD?
18:11 RealBadAngel was in hurry before sunset ;)
18:18 serengeor joined #minetest-dev
18:18 serengeor_ joined #minetest-dev
19:40 rsiska joined #minetest-dev
20:12 thexyz fuck this shit, how can irrlicht's core::list<MediaFetchThread> make win32 minetest to accessviolation with weird message when std::list<MediaFetchThread> works fine? am i doing something totally wrong?
20:24 SpeedProg joined #minetest-dev
20:35 jin_xi joined #minetest-dev
23:18 jin_xi joined #minetest-dev

| Channels | #minetest-dev index | Today | | Google Search | Plaintext