Time Nick Message
01:08 hmmmm back
01:24 hmmmm PilzAdam, check this out: http://ompldr.org/vaHc4Yg
01:26 hmmmm that's 8^3 -> 16^3
01:26 hmmmm ramp starting at 100, max height 600
01:48 hmmmm http://ompldr.org/vaHc4aQ 8^3->24^3, ramp starting at 0, max height 300
01:48 hmmmm to be honest i don't like this kind of ore
01:48 hmmmm can we just keep it the way it currently is?
01:53 ShadowNinja I like it, otherwise going beneath about -500 is pointless
02:56 VanessaE 22:56:05: ERROR[ServerThread]: register_ore: clust_scarcity and clust_num_oresmust be greater than 0ServerEnvironment::addActiveObjectRaw(): could not find block for storing id=10 statically (pos=(-241,16,-421))
02:56 VanessaE wtf?>
02:56 VanessaE (it did not cause a crash)
02:57 VanessaE I wonder which one triggered that "error"..
02:57 VanessaE perhaps this should be prefixed as "WARNING" instead?
03:26 VanessaE ah, just got that bad::alloc error
03:27 VanessaE http://pastebin.ubuntu.com/5651236/
03:27 VanessaE now with complete gdb backtrace.
03:28 VanessaE hmmmm: need anything else before I restart the server?
03:28 hmmmm excellent work
03:28 hmmmm no, nothing else
03:28 VanessaE ok
03:29 hmmmm the problem is clearly a Z coordinate of -50415
03:29 hmmmm who's doing what all the way out THERE!?
03:32 VanessaE HOW did anyone get there?
03:32 VanessaE oh wait
03:32 VanessaE I bet I know how
03:32 VanessaE I bet jeija's jumping mod sent someone out into nowheresville
03:33 VanessaE some obscure engine bug that happens with the right combination of trampolines.
03:35 hmmmm yeah. i can tell you more about the actual reason of the bug, just from looking at the backtrace (i don't know anything about collisionMoveSimple) i can see that there's a bad std::alloc, which means you probably tried to allocate too much memory, there's a really low Z, and i see min_z = -21978. putting two and two together, i suppose what happened is that it tried to allocate roughly cx * cy * 20000 objects and wasn't able to.
03:36 hmmmm i refuse to get involved with collision detection however, i can be more productive elsewhere..
03:38 VanessaE wouldn't it make sense to restrict players to within the normal range of the map?
03:39 hmmmm yes, of course. along with all entities.
03:41 VanessaE yeah
03:41 VanessaE I wonder if this is related to peaceful_npc...
04:09 VanessaE well at least now there something to show what happens and where.
06:57 celeron55_ < PilzAdam> we need some suggestions for a redesigned menu
06:58 celeron55_ this is very hard to do unless you limit yourself to irrlicht gui librarie's theming abilities (which basically mean you can switch the color of most things)
06:58 celeron55_ (but nothing else, and not even the color of everything, at least independently)
07:02 celeron55_ about collisionMoveSimple: it should work with anything; if it doesn't, sane limits should be added into it so that it doesn't try to allocate megabytes of stuff if something is wrong
07:03 celeron55_ but i don't see it allocating anything so i wonder what this is all about
07:04 hmmmm it's in a push_back()
07:04 celeron55_ oh, it does return an std::vector of collision events; maybe it tries to put too much stuff into that
07:04 celeron55_ should be simple to limit :P
07:05 celeron55_ oh and now i see more push_backs... these are probably the reason
07:06 celeron55_ and indeed it's most likely about the area that it calculates is needed for figuring out all collisions
07:06 celeron55_ it just needs a limit to speed_f and everything will be fine
07:07 VanessaE what would be the likely cause, from a userland perspective?
07:07 celeron55_ way too much acceleration
07:07 VanessaE thought so.
07:07 VanessaE jeija's jumping (trampolines) mod is the only one that could have triggered that
07:07 celeron55_ also, way too much speed will do to
07:08 celeron55_ too*
07:08 VanessaE *nod*
15:30 ShadowNinja Anyone with push access: https://github.com/minetest/minetest/pull/572
16:00 celeron55 urgh, the launchpad build would need curl and freetype added
16:01 celeron55 that means updating this file in this minetest-c55/packaging bazaar repo http://bazaar.launchpad.net/~minetestdevs/minetest-c55/packaging/view/head:/control
16:02 celeron55 and hell no, i'm not going to try that
16:02 celeron55 oh and, juhani isn't available at that e-mail address anymore and i don't have any other contact information of him; so he's gone
16:04 celeron55 i'll see if the guy who originally set up launchpad is around
16:06 celeron55 actually no, i'll not do that before launchpad has fixed the stable recipe thing
16:10 jojoa1997|Tablet is it possible to unload or override the default mod?
16:15 celeron55 jojoa1997|Tablet: you can override things in it, but not remove
16:16 jojoa1997|Tablet i want to override default so i can still load the stairs mod but not use default
16:16 celeron55 make a modified version of the stairs mod
16:17 jojoa1997|Tablet i tried copying all the mods into the game folder but then i got all unknown blocks
16:18 jojoa1997|Tablet brb
16:21 jojoa1997|Tablet back
16:22 jojoa1997|Tablet the thing is i have a modified verson of default and both are kinda merging properties
16:23 jojoa1997|Tablet i need to load default for dependency and also it stops me from having unknown blocks everywhere
16:23 jojoa1997|Tablet but i need to override it so only the modified version defines nodes
16:27 PilzAdam jojoa1997|Tablet, just add a empty default mod
16:27 PilzAdam and be sure to have proper aliases for everything
16:27 jojoa1997|Tablet pilzadam this id for your game minitest
16:27 PilzAdam I know
16:28 jojoa1997|Tablet i got a working version but it just aliases the prefix
16:28 jojoa1997|Tablet could i add this to the game config
16:28 jojoa1997|Tablet mods = minitest
16:29 PilzAdam I dont think this belongs to #minetest-dev
16:30 jojoa1997|Tablet sorry but i have been asking for a half in hour in the other channel
16:40 hmmmm so what did you think
16:40 hmmmm about the ores yesterday
17:11 hmmmm pilzadam, no comment?
17:11 PilzAdam hm?
17:11 hmmmm about the ores
17:12 PilzAdam you mean this: http://irc.minetest.ru/minetest-dev/2013-03-27#i_2959315 ?
17:13 hmmmm yeah
17:13 PilzAdam why dont you like it?
17:14 hmmmm we're still going to need differently-sized clusters and what not
17:14 hmmmm so we'd have to make just as many definitions
17:15 PilzAdam the current mapgen.lua has 3 register_ore calls for iron with the same size, and one with a bigger cluster size
17:15 PilzAdam so it is definetly useful
17:16 hmmmm i jsut checked and they're all different
17:16 PilzAdam also the smoothness of it is better than the current system
17:17 hmmmm it's not smooth though
17:17 hmmmm the scarcity difference only happens on chunk boundaries
17:17 hmmmm so if someone sets chunksize=15 or something, they won't see any difference in scarcity at all
17:17 PilzAdam well, its smoother than the current system
17:18 hmmmm i'm not so sure about that
17:18 hmmmm the first two transitions between scarcity and chunk size for iron happen well within a chunk
17:19 PilzAdam hm
17:19 hmmmm this whole thing sounded nice in concept but it doesn't look really good to be honest
17:20 hmmmm and it complicates things
17:20 hmmmm i'm all about doing more with less, and the less machinery we need to do accomplish the same task just as well, the better
17:26 PilzAdam what about this for gold: https://gist.github.com/PilzAdam/5256312 ?
17:28 hmmmm yeah looks good
17:28 PilzAdam Id also like to make mese more rare
17:28 hmmmm mese, or mese ore?
17:28 PilzAdam both
17:28 hmmmm okay, as long as it doesn't throw off the game balance
17:29 hmmmm why does gold get more rare the further you go down though?
17:29 hmmmm oh sorry, i read that wrong
17:29 PilzAdam how about 48^3 for the block, 16^3 and 12^3 for the mese ores?
17:29 hmmmm 48^3 is basically impossible to find
17:30 PilzAdam well, the block should be impossible to find
17:30 hmmmm well hold on a minute.
17:30 hmmmm a block is 80x80x80
17:30 hmmmm that means there are 4 of these per chunk
17:31 hmmmm *four*
17:31 celeron55 you wouldn't find it even with TNT 8)
17:31 celeron55 except by a rare-ish chance
17:32 celeron55 it could be desired though - it gives some dynamic to the game
17:32 PilzAdam the main source for mese are the ores
17:32 hmmmm that mese block better be the most awesome thing to exist, otherwise it's not worth finding them
17:34 PilzAdam you can still craft mese blocks out of 9 crystals
17:34 Exio what about mesecons?
17:34 Exio making them harder to obtain? why?
17:34 celeron55 by the way, i still have the "upstreamified" versions of the tnt (split to tnt and explosion) and the item pickup mods around that i made... well, somewhere like 2012-10
17:34 PilzAdam because mese tools are very powerful and shouldnt be too easy to get
17:34 celeron55 is there desire for those?
17:35 hmmmm back when I played minetest, I tried out the TNT and recall it being very buggy
17:35 PilzAdam hmmmm, wich TNT mod?
17:35 hmmmm don't really recall
17:35 Exio explosion in C++?
17:35 Exio celeron55: ^
17:35 celeron55 all in lua
17:36 celeron55 i added the one related thing to the lua api then though
17:36 hmmmm oh my god we need the voxelmanipulator in lua right NOW
17:36 PilzAdam celeron55, does it use the on_blast function?
17:36 celeron55 that one
17:36 hmmmm i'm tired with all the "let's put it in C++"
17:36 VanessaE for the logs if anyone cares, two crashes today... earlier, http://pastebin.ubuntu.com/5652641/ and just now, http://pastebin.ubuntu.com/5652875/
17:36 celeron55 i think TNT works just well as Lua-only
17:36 celeron55 the explosions are small enough
17:36 Exio i think there should be something, at least very bare, for defining the "hardness" of blocks
17:36 PilzAdam VanessaE, are you fine with mese being more rare?
17:37 celeron55 Exio: ehm... there is? digging time groups and levels
17:37 VanessaE PilzAdam: mese blocks you mean?
17:37 celeron55 they were designed with explosions in mind
17:37 PilzAdam VanessaE, and the deeper ore
17:37 Exio ah
17:37 VanessaE PilzAdam: yeah, it should be reasonably rare, and it isn't now, so go with what you've got and if it needs tweaked later, do that then
17:38 PilzAdam for the deeper ore: 9*9*9 -> 12*12*12
17:38 PilzAdam for the block: 16*16*16 ->48*48*48
17:39 VanessaE I'll have to take your word for it, I don't understand the new parameters yet
17:39 VanessaE but that *seems* fair
17:39 hmmmm how about 36
17:39 PilzAdam for the block?
17:39 hmmmm yes..
17:39 VanessaE blocks should be very rare
17:39 hmmmm yes, and they are
17:39 hmmmm you people have no effing idea how rare 36^3 is
17:40 VanessaE but not as much so as pop tart cats of course :)
17:40 hmmmm you think you know but you don't
17:40 VanessaE I have no f'ing idea how rare 9*9*9 is anyway
17:40 VanessaE :)
17:40 PilzAdam Ill make it 36^3 then
17:42 PilzAdam so, how about diamonds?
17:43 hmmmm diamond ore should be like 15^3
17:43 PilzAdam num_ores = 2?
17:43 PilzAdam so you cant make a pick if you find one clust
17:43 celeron55 thinking about the rareness of things is probably best done by thinking about how long worth of caves you are likely to need to walk in order to see it in the wall
17:43 hmmmm no way
17:44 celeron55 it goes roughly so that a single node length of cave will reveal roughly 12 nodes
17:44 hmmmm in minecraft, i think diamonds come in clusters of 4 or 5
17:44 PilzAdam lets make it 4 then
17:44 celeron55 thus, for something that appears once in 36**3 nodes, you need to walk 36**3/12=3888 nodes worth of cave
17:45 celeron55 that's further divided by the size of a cluster though
17:45 PilzAdam wich min height? -1024?
17:45 hmmmm no that's insane
17:45 hmmmm pilz your judgement blows
17:45 hmmmm should start at -32 say
17:45 celeron55 go play some minecraft alpha; it's lovelily crowded
17:46 PilzAdam hmmmm, mese starts at -64
17:46 hmmmm it's not mese though
17:46 PilzAdam diamonds are better than mese
17:46 PilzAdam so they should be a bit more rare
17:46 hmmmm hrmm
17:46 hmmmm alright, i guess we can start diamonds at -128
17:46 hmmmm and you use the mese to get to diamonds or something
17:46 celeron55 i think the availability of an ore should not be singlely dependent on how frustratingly deep it is
17:47 hmmmm got anything better/
17:47 celeron55 there should be something in it that makes it challenging to get other than just the deepness
17:47 celeron55 i don't have anything pratical, only completely untested ideas
17:47 hmmmm well what are the idea? i can probably shoot a couple down 8)
17:48 hmmmm s/idea/ideas/
17:48 celeron55 sure
17:48 PilzAdam so, I have scarcity=14^3, num_ores = 4 and height_max=-128
17:48 celeron55 at least one is that the tools for mining an ore (or a layer that surrounds it) are obtainable only from somewhere else than the ore itself
17:48 hmmmm hrmm, seems good pilz
17:49 PilzAdam celeron55, currently you need a steel axe to get mese and diamonds
17:49 PilzAdam *pick
17:49 celeron55 it's a lighter implementation of that idea, yes; but quite light and single-dimensional
17:49 hmmmm i think what he means is like
17:50 hmmmm maybe you need an enchantment on a tool in order to mine some super rare and good stuff
17:50 hmmmm or have some sort of item there are only a handful of at any given point on the map
17:50 hmmmm that would be a cool idea
17:51 celeron55 or that are obtainable only by fighting a certain monster, or something
17:51 hmmmm yeah but that necessitates having mobs
17:51 celeron55 yes, i know; that is why i left it last 8)
17:51 PilzAdam https://github.com/PilzAdam/common/compare/master...ores
17:52 hmmmm anyway i was thinking maybe there can be a single really epic tool that is generated once on the map and never again after that
17:52 hmmmm and it's left on the surface
17:52 hmmmm not underground or anything
17:52 celeron55 if there is something like that, then there should be similarly globally generate clues for where it is, because otherwise it's impossible to find
17:52 celeron55 generated*
17:53 PilzAdam VanessaE, are you ok with this: https://github.com/PilzAdam/common/compare/master...ores ?
17:53 celeron55 this is actually something for which on_generated can be useful - very game-specific, and only done in rare places for a small amount of nodes
17:53 PilzAdam the aliases for moreores are added in minetest_game
17:53 celeron55 (maybe)
17:53 hmmmm well the thing is, it wouldn't be a 1 in (31000/80)^2 chance like you might think, it could a much lower chance so the user finds it faster
17:53 hmmmm much higher chance*
17:54 celeron55 then there is the multiplayer aspect which screws up all nice singleplayer game thinking
17:54 celeron55 8)
17:54 hmmmm and again, because there's a lock for only one at any given time, a higher probability of finding it won't mean there are more of them
17:54 hmmmm well no that's fine
17:54 hmmmm because look, conflict
17:54 hmmmm "who controls the epicpick"
17:55 hmmmm and then people can fight eachother for the pick
17:55 VanessaE I thought the plan was to put all the new ores in all at once?
17:55 celeron55 should we design minetest mainly so that the gamey aspect is made for singleplayer, and people can tune their servers around it as they like
17:55 PilzAdam or someone gets the epicpick and leaves the server forever?
17:55 hmmmm vanessa, that was the plan, yeah.
17:55 VanessaE otherwise, this all looks okay to me, save for the question of how rare mese blocks should be
17:56 VanessaE hmmmm: ok, just making sure I didn't miss a step.
17:56 hmmmm pilzadam, that's the thing - have to iron that detail out, but i'd say make sure it can't be put into a locked chest
17:56 PilzAdam why put all ores in at once?
17:56 celeron55 a finished game is so much about finishing frustrating details that anyone who hasn't done it won't believe it
17:56 celeron55 minetest has never gone through that stage
17:57 VanessaE PilzAdam: do it once, do it right?
17:57 PilzAdam what ores do you want?
17:58 hmmmm i don't think it really matters
17:58 hmmmm as long as they're all there before a release
17:58 PilzAdam I mean, diamonds and gold are ready yet; why should we wait?
17:58 hmmmm celeron, that was just one idea though, what were the others?
17:59 hmmmm and i think it isn't too difficult to implement either
18:00 PilzAdam hmmmm, I had a random thought about the cave air replacing leaves: maybe make it only replace things that have ground_content = true in nodedef?
18:00 hmmmm you know, here's the funny thing: i've been treating minetest like it was just some program and not a game that people play. that is, backward compatibility is of the utmost importance, there are 100 things to configure, i even refer to people as "users" instead of "players"
18:01 hmmmm pilzadam, now there's a good idea
18:01 hmmmm why is it that i never consider the node properties, just the specific nodes themselves
18:02 celeron55 well, then, making some kind of environmental effects to eg. turn ore or something into an another ore... a bit like an extension to lava+water making cobble and obsidian; maybe causing some pressure on some ore, or letting hot air out from some type of nodes from the place where they originally are underground
18:03 PilzAdam celeron55, why was ground_content added to the default nodes?
18:03 celeron55 and this i mean on a natural node level, not some pumps and pipes with text fields
18:03 hmmmm the former example could be done with claylike ore perhaps
18:03 celeron55 PilzAdam: it's a legacy think
18:03 celeron55 thing*
18:04 PilzAdam so, can we re-use it?
18:04 celeron55 it once was some kind of a hint for the map generator
18:04 celeron55 you can, but be aware that mods use it very haphazardly
18:04 celeron55 as it's in some examples and actually does nothing
18:05 hmmmm so what do we call the legendary pickaxe there's only one of? :p
18:05 hmmmm i vote for "Epick"
18:05 PilzAdam hmmmm, so, lets give dirt, dirt_with_grass, stone and sand ground_content=true and let the mapgenerator only replace them?
18:05 hmmmm pilzadam, yes
18:05 hmmmm also all ores need to be ground_content=true
18:05 PilzAdam and add it to lua-api.txt
18:06 PilzAdam actually its called is_ground_content currently; should we rename it?
18:06 hmmmm *shrug*
18:07 celeron55 i think having "is" or "has" or some other on/off verb at the beginning of boolean values is good
18:07 celeron55 for "ground_content", you'd think that it would define "what" the ground contains
18:07 PilzAdam ok
18:08 PilzAdam what about sandstone?
18:10 hmmmm sandstone too of course
18:10 hmmmm don't forget cobble either
18:10 hmmmm need to be able to let caves eat dungeons
18:10 hmmmm and gravel
18:10 PilzAdam https://github.com/PilzAdam/common/commit/1c0c6a8db604fca740eda38e021a99b35dfafcf1
18:11 PilzAdam every solid node had it already
18:11 celeron55 i guess the majority of things are "ground content" and the minority is things that shouldn't be eaten by the map generator
18:11 VanessaE shouldn't only those things added by the default mapgen be in that group in the first place?
18:11 celeron55 maybe the reverse would be better? 8)
18:12 VanessaE (minus trees/leaves)
18:12 PilzAdam VanessaE, no
18:12 PilzAdam that would require a change of the mods if the mapgen changes
18:12 VanessaE hrm
18:12 VanessaE I guess so.
18:13 PilzAdam what about stairs?
18:13 hmmmm stairs too yes
18:13 celeron55 basically only those should have is_ground_content=false that are put in the world at world generation but shouldn't be eaten by caves
18:14 celeron55 that means trees and... ehm... what else?
18:14 PilzAdam papyrus, cactus, grass
18:14 celeron55 small things should be, because otherwise they'll be floating in air without any support while the cave generator would have eaten them nicely completely anyway
18:15 celeron55 trees are a problem only because they can be larger than caves and be left partially exist
18:15 celeron55 so tree-like things (eg. generated buildings) are such
18:16 celeron55 this leads to a funny thing
18:16 celeron55 stairs would not be in that one because of dungeons, but still if there was stairs in a generated house, they shouldn't be removed by caves
18:17 hmmmm but generated houses would be made out of other ground content too
18:17 celeron55 so you think generated houses shouldn't work like trees?
18:18 celeron55 why do trees need to have special behavior anyway
18:18 PilzAdam what about liquids?
18:18 hmmmm that whole thing is very difficult
18:18 celeron55 if buildings don't need to
18:18 hmmmm liquids, false
18:18 celeron55 i'd imagine having a cave go through a building is worse than having a cave go through a large tree
18:18 hmmmm the cave generator already does a special check to not replace water or lava
18:18 hmmmm hrmmm this is such a difficult problem
18:19 celeron55 i think you should just ignore it; it's too much work for too little benefit and every solution has a downside 8)
18:19 hmmmm i mean
18:19 hmmmm it's not what was placed
18:19 hmmmm instead, it's how it's been placed
18:19 celeron55 it's all about the context, yes
18:19 celeron55 vmanip is able to preserve context via flags
18:19 hmmmm the obvious solution would be to store a flag to not replace this
18:20 hmmmm but we don't have a bit to spare
18:20 celeron55 but they don't stay between generator invocations
18:20 celeron55 well, you often have
18:21 celeron55 but eg. for stairs, you don't
18:21 celeron55 they use lighting, and they use facedir
18:21 hmmmm erhm, would you like to take it out of the content?
18:21 hmmmm i kid.
18:21 celeron55 take what out of... where?
18:21 hmmmm the content field
18:21 hmmmm i was joking, i don't really think we should do that.
18:22 PilzAdam https://github.com/PilzAdam/common/commit/58927129aa194be0f10ee9fd283fe66c8e670696
18:22 celeron55 it's arguable that the id space is too small already
18:22 hmmmm 16 bits is crazy
18:22 celeron55 people have crazy amounts of nodes
18:22 hmmmm people are going to really have 65000 nodes? really
18:22 hmmmm 65000 is nuts though
18:23 celeron55 go tell VanessaE, i think she was the first to go over ~2048 that was once the limit
18:23 hmmmm also a lot of those nodes are duplicate ones with different hues
18:23 celeron55 and that was LONG ago
18:23 hmmmm imagine all the definitions needed for 65000 nodes
18:23 celeron55 i already said it's completely crazy; but that's how it is 8)
18:24 hmmmm anyway with the 6d facedir that will probably be cut back
18:25 VanessaE my server has some 2600 defined nodeas
18:25 hmmmm certifiably insane
18:25 celeron55 have you used each of them? 8)
18:26 VanessaE each mod on my server has been used, yes.
18:26 VanessaE can't exactly pick and choose one node from here, two from there :)
18:27 VanessaE (except, I guess no one uses mese starter kit, but PilzAdam is about to integrate that into default anyway)
18:27 ShadowNinja the meta nodedef branch should cut the number a lot more
18:27 VanessaE ShadowNinja: 6d facedir will help too, eventually
18:27 PilzAdam what keeps us from merging this branch in right now: https://github.com/PilzAdam/common/compare/master...ores ?
18:28 hmmmm PilzAdam, nothing
18:29 ShadowNinja you raised the digging times of the mese pick!
18:29 PilzAdam yes
18:29 PilzAdam because we have diamonds now
18:29 PilzAdam and diamonds are a tiny bit faster and last longer than mese tools
18:30 VanessaE as long as mese pick is still faster than steel, I don't see any reason it can't be a little slower than a diamond one
18:30 ShadowNinja well diamond should be faster than it is
18:30 PilzAdam faster would be instant dig
18:30 hmmmm heh, imagine how ridiculous things would be with efficiency 5
18:31 hmmmm like netherrack
18:31 PilzAdam btw: pushed
18:31 VanessaE ever used an admin pick? :-)
18:31 VanessaE THAT is fast. :)\
18:31 ShadowNinja is it realisticaly possible to use only diamond picks or would you run out before you were able to get a new one?
18:32 ShadowNinja infinitytools pick, I use that on my server
18:32 hmmmm no clue.. people should be able to do grid mining for diamond and mese
18:32 VanessaE imho, using a diamond pick should be just a *hair* less than sustainable
18:32 VanessaE force the player to not waste it on cutting stone]
18:33 ShadowNinja I think it should be more, that way you have diamonds for things other than just making annother pick
18:33 hmmmm indeed
18:33 celeron55 hmm, let me see
18:33 VanessaE (as in for every say 25 diamonds you collect, you have to use 30 to keep making diamond picks)
18:34 hmmmm yeah but vanessa, that's one thing that bugged me
18:34 hmmmm that sort of unsustainability should force you to go deeper where the ores are more common
18:34 VanessaE p-erhaps
18:34 VanessaE --
18:35 hmmmm because otherwise there's no real point to mining
18:35 VanessaE when I last mined for any length of time, mese was common enough to be self-sustaining + build up a nice cache of the stuff
18:35 celeron55 the diamond pick will have 810 uses mining default:stone, and by mining 2 of those, you see 6 new nodes that could be diamond
18:36 celeron55 and there are 3 diamonds in 14**3 nodes
18:36 celeron55 and you need 3 for a pick
18:36 celeron55 it means that no, it's not worth of grinding
18:37 celeron55 by a factor of 3.4
18:37 hmmmm yeah exactly
18:37 hmmmm but again, it might also force people to come up with smarter ways to grind
18:37 hmmmm like i said, grid mining
18:38 celeron55 grid mining?
18:38 hmmmm you mine in a grid of 3-high slots spaced by 2 or 3 blocks apart
18:38 celeron55 hmm, actually, i wonder how much that can be optimized by knowing that diamonds are in clusters... it probably means that the gain is effectively multiplied by the cluster size
18:39 celeron55 that'd get it to 3.4/3=1.13
18:39 hmmmm that way you'd be able to tell if there's a cluster of ores and dig the rest out, and if not, you'd save a whole bunch of time and pick durability by not mining those
18:39 celeron55 by the other stuff that you get, it'd keep it worthy of doing
18:39 celeron55 just barely
18:40 celeron55 i guess it's one way to balance it, altough not sure if it's a good way 8)
18:40 hmmmm diamonds might have to be a tad more common.. it's too early to tell though
18:40 PilzAdam random: I took a diamond axe and mined straight down; 30 secs and I found mese
18:40 hmmmm we'll see, based on how many angry people there are on the forums and irc
18:41 celeron55 hmmmm: good approach
18:41 celeron55 if there aren't enough, it must be made rarer!
18:41 hmmmm it's funny how angry people get over a completely free game that's just a game
18:41 celeron55 a game is good only if it makes a portion of people angry 8)))
18:43 PilzAdam ok, lets make celeron55 angry
18:43 hmmmm completely unrelated
18:43 hmmmm what if the NodeIdMapping was sent independent of the blocks
18:44 hmmmm NameIdMapping rather
18:44 celeron55 sent? i don't think it's sent
18:45 hmmmm ahh it's not
18:45 hmmmm that's only for disk
18:45 hmmmm smart move
18:49 hmmmm but anyway i think we should focus more on loading from disk times
18:50 hmmmm IIRC it takes 30ms to load a block from disk, and the map generation takes about 70ms total... for a chunk... which is 25 blocks
18:57 PilzAdam I think mese is too common
18:58 PilzAdam so either we make it more rare or we make the tools slower
18:58 PilzAdam VanessaE?
18:58 VanessaE more rare I think
18:58 VanessaE no one likes slow tools
18:59 VanessaE but not too much moreso; didn't you already do that with that last commit anyway?
18:59 PilzAdam its still too common
18:59 VanessaE hrm
18:59 PilzAdam and I only modified the bottom layer
18:59 VanessaE even at 36A^3?
18:59 VanessaE er 36^3
18:59 PilzAdam Im not talking about the blocks
19:01 ShadowNinja hmmmm: will ores registered with register_ore() become more common as you go downward?
19:01 PilzAdam no, we dropped this idea
19:01 VanessaE oh you meant the ore
19:01 VanessaE um
19:01 ShadowNinja :-(
19:02 VanessaE how do the settings in the new ore method compare to the old default.generate_ore() call?
19:02 VanessaE about the same as then?
19:02 PilzAdam ShadowNinja, it sounds nice in theory, but in practice you dont notice it
19:02 hmmmm shadowninja, nope
19:02 PilzAdam VanessaE, yes, the same
19:03 ShadowNinja I noticed it in 0.3, lots of mese at -30900
19:03 VanessaE ok, I would say in that case, make ore cluster together closer, increase that setting by maybe 1, and make the overall concentration about half as dense per chunk as it currently is
19:03 VanessaE (that setting = the cluster size)
19:03 PilzAdam ShadowNinja, currently we use different calls of register_ore() to make it more common deeper in the ground
19:03 VanessaE so you get larger, denser clusters, but there are far fewer than there were before
19:04 ShadowNinja And players have gond down that far in "survival", I have personaly been to -6000
19:04 VanessaE ShadowNinja: lowest I've dug in survival is about 7k
19:04 celeron55 hmmmm: i just ran minetest for a while in callgrind and opened this in kcachegrind
19:05 ShadowNinja I know, I think there should be an int for the increase, that way if you don't like it you can just set it to 0
19:06 celeron55 i've done this before too, but 1) this is very messy to browse and 2) it's very hard to try to gather any meaningful data from this
19:06 RealBadAngel about last error reports. i suggest capping entities both on acceleration and speed. and reverting entity being on the loose
19:07 PilzAdam VanessaE, currently the upper layer has 3 ores/clust with a scarcity of 16^3 and ranges from -64 to -127
19:07 RealBadAngel terminal velocity and max acc value
19:07 jojoa1997|Tablet entities should stay near where they spawned unless players lead them away
19:07 PilzAdam VanessaE, Id say increase scarcity to 18^3 and make it down to -257
19:08 PilzAdam and for the deeper layer, increase scarcity to 14^3
19:08 RealBadAngel celeron55, you have noticed z value being 50k but missed x being 209k
19:09 celeron55 RealBadAngel: you don't need to limit acceleration if you limit speed
19:09 RealBadAngel true
19:09 celeron55 also, there isn't need to limit anything anywhere else than inside collisionMoveSimple
19:10 PilzAdam what about calling the speed limit the "speed of light"?
19:10 RealBadAngel terminal velocity just
19:10 hmmmm ShadowNinja, if you want it, here http://pastebin.com/z6pUy2FR
19:10 VanessaE [03-27 15:08] so lets cap it at 500 m/sec - which is fucking insane speeds, but acceptable for a video game
19:10 VanessaE [03-27 15:08] so that would be...\
19:10 VanessaE [03-27 15:09] 51 seconds.
19:10 VanessaE [03-27 15:10] to reach 500 m/sec if you start at 0 and fall at 9.81 m/s/s
19:10 VanessaE there, to end that discussion before it starts L()
19:10 VanessaE :)
19:11 VanessaE (in reality, we already know a human can exceed 340 m/sec if specially equipped)
19:11 RealBadAngel VanessaE, we are not really talkin bout physics now
19:11 VanessaE RealBadAngel: of course not, but we're talking about acceleration and speed.
19:12 VanessaE both of which should be capped
19:12 RealBadAngel we are talkin bout entity bein unatended and out of control, thus having weird pos and speeds
19:12 celeron55 there is no need to cap acceleration; stop saying it
19:12 RealBadAngel because server has not responded
19:13 hmmmm hah, reminds me of Big Rigs
19:13 hmmmm no acceleration cap on reverse
19:13 Calinou g_bugrigs 1
19:13 celeron55 hmmmm: correction: speed cap
19:13 RealBadAngel i had same problem with tubed items
19:13 celeron55 acceleration is a static value in it, = capped
19:14 celeron55 but that is irrelevant in this case
19:14 RealBadAngel i capped them with last_valid_position and demanded them to report their next
19:14 hmmmm i meant speed cap actually
19:15 RealBadAngel if they fail to report, theyre rolled back to last_valid
19:21 celeron55 shit crap fuck, i can't connect with a regular client to a minetest server running under callgrind because "Server: peer_id=2: failed to emerge player"
19:21 celeron55 this is 100% reproducable and predictable
19:22 celeron55 why the hell
19:23 celeron55 ...i get this without callgrind too
19:23 celeron55 this isn't even fair
19:24 celeron55 what?!
19:25 celeron55 now it works
19:25 celeron55 fully
19:25 VanessaE bbl
19:36 celeron55 hmmmm: i think you might have wrong presuppositions about the need for a voxelmanipulator for lua
19:36 celeron55 hmm
19:37 celeron55 actually no, this wasn't what i thought this was
19:38 RealBadAngel celeron55, lately i supposed that improper destructors was the cause of bad alloc, made some patches, restarted game for like 20 times, then exited to OS without any error
19:39 RealBadAngel then to test reverted the changes, it behaved same
19:39 RealBadAngel next run with pathces it crashed 2nd try
19:39 sapier :-) good test ... not quite surprising as this is exactly what gcc warning predicts
19:40 celeron55 well, i guess the standard allows it to make your computer monitor pull out a gun and shoot you
19:41 celeron55 hmmmm should seriosuly fix those missing virtual destructors
19:41 celeron55 seriously*
19:41 sapier I doubt this is a standard problem c/c++ allows almost everything and expects programmer to have enough knowledge to decide what's best for her/him
19:42 celeron55 they're likely to not cause anything else than memory leaks but that's too bad already
19:43 sapier I've almost completed scriptapi rewrite I'd like to give an early preview prior fixing all formal issues what's best way to get comments on it?
19:43 celeron55 anyways, i can tell some interesting things from this callgrind's output of a short run of a server with a player walking around making the map generate and load
19:43 RealBadAngel like what?
19:43 celeron55 for example, the vast main CPU hog of the default ABMs is l_find_node_near
19:44 celeron55 which sounds like leafdecay to me
19:44 sapier not only ... most modders use this function for spawning abms too
19:45 RealBadAngel abms are heavily lagged lately
19:45 celeron55 and the ABMs take, mainly because of that, most half of the server environment step time
19:45 celeron55 -half
19:45 sapier true that's why most of moders try to avoid abms and do things like that at on_generate
19:46 celeron55 i can also tell why l_find_node_near is so heavy: CNodeDefManager::getIds() takes the vast majority of it's execution time
19:46 celeron55 the result of commonly used inputs to that could probably be cached
19:47 RealBadAngel first of all abms are a single thread. and theyre able to stop whole system
19:47 sapier maybe using some sort of hash table
19:47 celeron55 the problem here are not ABMs
19:47 celeron55 stop misinterpreting me!
19:47 celeron55 ABM is just what calls this, it is not the problem
19:47 sapier it's the way abms are done right now
19:47 celeron55 ...
19:48 celeron55 okay, i'll just say nothing as you don't understand anyway
19:48 RealBadAngel try to explain
19:48 celeron55 i already explained; it's your turn now
19:48 celeron55 your turn to understand
19:48 sapier is my interpretation correct celeron?
19:48 celeron55 NO.
19:49 RealBadAngel i said that a few months ago. with a single ABM call, a loop i can take over whole ABMS system
19:49 RealBadAngel and stop it
19:50 RealBadAngel now your turn to say it is ok
19:50 hmmmm lua just blows, not really news
19:50 sapier it's not exactly lua
19:50 celeron55 now hmmmm is completely wrong too :D
19:50 celeron55 this is hilarious and i can't even think of any words that would make RealBadAngel or sapier any more right
19:51 hmmmm what are we talking about?
19:51 celeron55 i'll go to sauna now, maybe you'll get even more wrong while i'm away ->
19:51 RealBadAngel hehehe
19:51 hmmmm that's just so wrong
19:52 sapier hmm ... he said abms call somethin internal which is taking very much time ... is that true or am I already wrong with this?
19:52 RealBadAngel abms are made one by one, in a line
19:53 RealBadAngel if one is takin too much time rest of them waits
19:53 hmmmm hash_node_position? what's that
19:53 sapier yes but if it's taking to much time on a single core machine you can't do anything against this
19:53 RealBadAngel thus if you decide to make infinite loop in abm
19:54 RealBadAngel you took over the matrix
19:54 hmmmm realbadangel, same with any mod
19:54 hmmmm that's just the way it is
19:54 RealBadAngel no more abms
19:54 sapier and this will happen at any case in any lua mod
19:54 hmmmm the only find_node_near that i see is the one that gets ignore netx to tree
19:54 hmmmm and it's not in a loop at all
19:55 hmmmm so maybe find_node_near is really that crappy
19:55 hmmmm let's see
19:55 RealBadAngel we are talkin here bout both complex and badly written mods
19:55 RealBadAngel effect is the same
19:55 hmmmm ? we're talking about leaf decay
19:56 RealBadAngel after 10 seconds waiting for stargate abm teleport code you are able to hear 10 teleporting sounds in a row
19:56 sapier CNodeDefManager::getIds() is this done for every node?
19:56 hmmmm yes? where
19:57 RealBadAngel hmmmm, and leaf decay is also known killer
19:59 hmmmm hrmm i wouldn't expect it to be slow
19:59 hmmmm it doesn't seem like it really does that much
19:59 hmmmm woah woah okay
19:59 hmmmm nevermind
19:59 hmmmm getFacePositions looks bad
20:00 hmmmm perhaps the use of a manualmapvoxelmanipulator is warranted here
20:00 sapier lol
20:01 sapier we're looking at any nodes around of nodes within range
20:01 hmmmm well i'm not really worrying about this at the moment
20:01 sapier thus checking many nodes multiple times?
20:02 sapier first we check them for radius 1
20:02 sapier then for radios 2
20:02 sapier then for 3 ...
20:02 RealBadAngel i do have my own problem, rotating vertices for texturing seems to rotate the lightnin too :(
20:02 hmmmm heh
20:02 hmmmm that's a minor cosmetic problem
20:02 hmmmm i wouldn't worry too much about that
20:02 RealBadAngel but have to be solved
20:03 sapier we should remove the for loop from find node near and it'l be lot faster
20:04 sapier hmm wait ... seem's I've misinterpreted the code again
20:05 hmmmm i feel like we can just redo that entire loop and it'd be way better
20:06 RealBadAngel why cant we have this actualised on action?
20:06 sapier I'm not quite sure why it's that bad ... it's looking for first node starting from in to out ... this is quite obvious from real life but maybe is worst case scenario considering memory allocation
20:07 RealBadAngel like gravel in the air when touched fallin
20:07 hmmmm who cares if it's slow, right guys?
20:07 hmmmm just buy a faster computer!
20:07 RealBadAngel hehe
20:08 RealBadAngel i say dont care bout leaves. until trunk is cut?
20:08 hmmmm i'd say so too but that's hard
20:08 sapier it's not leaves only
20:09 RealBadAngel but a case
20:09 RealBadAngel dont do, until youre forced to do so
20:10 RealBadAngel instead of doing it all the time and consuming power
20:10 sapier still celeron told the getIds is consuming most of cpu power so the loop isn't worst thing in there
20:11 sapier here we go: for(u16 id=0; id<=MAX_CONTENT; id++)
20:12 sapier max content is 0xfff ... this loop is done completely everytime
20:12 hmmmm now let's see here
20:12 hmmmm what does getIds actually do
20:13 hmmmm returns a set (why a set?) of IDs given some names?
20:13 celeron55 <-
20:13 hmmmm oh, given a group name
20:13 celeron55 the only thing i said and the only thing i know is that getIds takes most of time when default ABMs are run
20:13 RealBadAngel so fast back from a sauna?? water was cold or what? ;)
20:14 hmmmm doesn't matter.. looping 65000 times is ridiculous and is POOR QUALITY CODE.
20:14 hmmmm this needs to be fixed.
20:14 sapier we don't have found the solution yet go away ;-)
20:14 hmmmm obviously the solution is to find an actual solution to matching IDs with a group name
20:14 hmmmm not just searching everything
20:15 hmmmm what's the problem with having a std::map of group names-> vector of IDs?
20:15 sapier it's only 4096 times but bad enough
20:15 celeron55 hmmmm: no problem, it just hasn't been done
20:16 celeron55 this is really again a proof-of-concept implementation that was forgotten
20:17 hmmmm have you no shame of releasing this code to the general public!?
20:17 PilzAdam we should have a "proof-of-concept implementation" somewhere
20:17 PilzAdam +list
20:17 celeron55 hmmmm: not really; i'd have some shame if i asked money for it 8)
20:17 sapier I have already a complete list: minetest ;-)))
20:17 celeron55 :D
20:17 hmmmm you're right. the buildaworld guys are right too. scrap it and start all over
20:18 hmmmm soooooo
20:18 hmmmm who wants to take care of this
20:18 * RealBadAngel takes care of lights....
20:19 sapier :) I doubt anyone of us (except celeron) has already done enough mistakes to avoid them on rewrite :-) so we'll do them when trying to rewrite it
20:19 celeron55 anyways, maybe i'll finish the implementation
20:19 celeron55 i wonder what else i'll find by using this glorious tool of callgrind
20:20 hmmmm why the framerate is lower than it was in older versions
20:20 celeron55 i gave you two suggestions when you last asked
20:20 hmmmm i disabled shaders and disabled fancy leaves, that didn't give me all my fps back
20:20 hmmmm what were they?
20:20 RealBadAngel hold on, last 2-3 weeks ive noticed im hittin 60fps again
20:20 celeron55 hmmmm: shaders and non-d3d
20:21 celeron55 i don't know if you run windows on that
20:21 hmmmm i don't
20:21 celeron55 then i have no idea and you'll have to find yourself
20:21 hmmmm i used to think that FPS didn't make too big of a difference above 30, but it really does
20:21 RealBadAngel 6d facedir alone freed some of the memory and speeded up the game
20:21 hmmmm the game just doesn't feel as smooth
20:22 celeron55 RealBadAngel: only if you run mods that define craploads of nodes otherwise (and use HD textures)
20:22 hmmmm i was trying out some older 0.4-dev versions to see if i can match up with the map generator of my first world (i couldn't), and i just couldn't believe how much faster the graphics were
20:22 celeron55 for default, it doesn't make any difference
20:22 RealBadAngel 400mb in case of vanessae server
20:23 RealBadAngel celeron55, you shall said "not developed worlds"
20:24 hmmmm huh
20:24 RealBadAngel any facedired node tripled the textures before
20:24 RealBadAngel because of transform usage
20:26 RealBadAngel if with as you said, suprisingly small code i earned 400mb for a server, it was worthy, dont you think so?
20:26 celeron55 hmmmm: anyway, if you find the change that matters for you, it'll probably matter for some others too
20:26 hmmmm eh yeah
20:27 celeron55 graphics performance is a bitch though
20:27 celeron55 different computers have grossly different speeds for doing different things
20:27 celeron55 eg. some might be extremely slow on texture changes
20:27 celeron55 some might choke on mesh uploads to the GPU
20:27 celeron55 and so on
20:27 RealBadAngel for me 6d facedir effected circa +10% fps gain
20:27 hmmmm yeah but you use huge textures
20:27 RealBadAngel 128x
20:27 hmmmm ya huge
20:28 RealBadAngel but cmon
20:28 RealBadAngel its 4 yrs old box
20:28 RealBadAngel nothin modern
20:29 RealBadAngel 1,8ghz dual core with 4gb
20:29 hmmmm i couldn't get minetest to run hardly at all on my laptop when i tried it out last week
20:30 RealBadAngel btw i noticed great speedup lately
20:30 hmmmm what i wonder is why people didn't catch regressions before they got old
20:30 RealBadAngel i moved all the ore generation to new function
20:30 RealBadAngel even not my own ores
20:30 RealBadAngel and starting new world takes a second now
20:31 celeron55 hmmmm: i guess such regressions aren't considerable enough
20:32 celeron55 but seriously, it would be interesting to know what did that to your laptop
20:32 hmmmm i need to stop dicking around and get to work on this
20:32 RealBadAngel hmmm the regressions in our case are imho because we do more, want more and expecting more
20:33 RealBadAngel thus engine gives up
20:33 hmmmm the thing is
20:33 hmmmm compiling takes so much longer for my laptop
20:33 hmmmm i can't cross compile since that's using older libraries than my desktop and it probably won't link right
20:35 Exio what about a chroot? or -static?
20:35 Exio (yes, very good answers)
20:35 RealBadAngel may i offtopic a bit?
20:36 RealBadAngel https://github.com/minetest/minetest/blob/master/src/game.cpp#L846
20:37 RealBadAngel can we rename those files and make local music?
20:38 hmmmm exio, eh..
20:38 hmmmm ahh i need to try 0.3.x again on my laptop to see what the actual performance is
20:38 hmmmm i forgot about how i upgraded to 310 drivers on my laptop
20:38 hmmmm which caused a regression for half life 2
20:40 RealBadAngel hmmmm, have u seen the code?
20:41 hmmmm yeah, i don't really have an opinion on it
20:41 RealBadAngel i say time for main menu music score at least :)
20:41 hmmmm i don't need background music because i have a music player, so does everybody else
20:41 hmmmm and it does the job way better
20:42 celeron55 i don't think we can decide on any music to distribute officially with the engine (not even with the distributed games)
20:42 celeron55 so it's kind of a problem
20:42 RealBadAngel well, technic ambience tracks are allowed for us to use
20:43 RealBadAngel we got the permission
20:45 RealBadAngel http://realbadangel.pl/technicambience.zip
20:46 VanessaE back.
20:46 RealBadAngel wb
20:52 celeron55 i actually compiled and booted 0.3.3 now
20:52 celeron55 there is something in this that makes it awesome
20:52 celeron55 i don't fully understand what
20:52 Exio the mapgen!
20:52 * Exio looks at hmmmm
20:54 RealBadAngel celeron55, its nostalgy. done and gone ;)
20:55 celeron55 no
20:55 celeron55 i see this as a different kind of game
20:55 sapier https://github.com/sapier/minetest/commit/32a76041cf34cd650b06a1613abf085f9ccf1e19 this is a complete rewrite of scriptapi integration ... it's an early version not ready for merge as comments file headers, filenames etc aren't ready but it should already be functional equivalent to current scriptapi
20:55 RealBadAngel celeron55, ofc. project is evolving
20:55 celeron55 this is kind of what you would expect from a game, but is missing a few details; 0.4 is kind of... not
20:56 sapier maybe someone is interested in commenting and giving hints what needs to be fixed ... main goal was to reduce header linkage and separate scriptapi lua parts from core
20:57 RealBadAngel celeron55, you propably seen mc indev and later
20:57 celeron55 0.4 has it's place in this world, but i'm constatly feeling more so that 0.3 has to be revived as a fork
20:57 celeron55 it has value on it's own
20:58 PilzAdam it would be better to focus on one game instead of forking it
20:58 RealBadAngel celeron55, if it is lightweight maybe fork it on other systems like portables
20:58 celeron55 it's just like how i use MATE as the desktop environment at work - a gnome 2 fork
20:58 celeron55 some probably said gnome 2 should die and everybody should focus on 3
20:59 celeron55 but at the end, it wasn't a good idea and the world is happier with both
20:59 RealBadAngel see, im Might and Magic player
20:59 RealBadAngel its 6th out there
20:59 RealBadAngel and im still playin the 3rd only
21:00 hmmmm but what's different about 0.3.x?
21:00 RealBadAngel which was out more than 15yrs ago
21:00 PilzAdam hmmmm, the textures
21:00 hmmmm is it really just the mapgen?
21:00 hmmmm there are so many technical improvements with 0.4.x
21:00 hmmmm from a rational standpoint you'd have to be crazy to like 0.3.x better
21:00 sapier1 :-)
21:00 RealBadAngel it is lack of contens
21:01 RealBadAngel thus it is not very demanding
21:01 celeron55 0.3 is minimalistic
21:01 celeron55 and it has a mapgen that makes very good use of it's small amount of node types
21:01 RealBadAngel and then could be moved in such state to android or whatever
21:01 sapier1 everything has use for someone
21:02 hmmmm well let's bring back mapgen v5
21:02 celeron55 and it has mobs that make some sense and work
21:02 Exio add sapier1's path finding to upstream
21:02 RealBadAngel 5??? gimme biomes :)
21:02 RealBadAngel i want SEVEN
21:03 hmmmm mmm
21:03 hmmmm see, that's the thing
21:03 hmmmm i want to work on something new
21:03 celeron55 have you tried booting up 0.3? it takes 0 time units because the mapgen works on a block-by-block basis and it doesn't need to check what content the server uses, it's always the same
21:03 hmmmm but we also have to keep going back and fixing all these other bits of nonsense
21:04 sapier1 of course simple things can be done fast but complex things can be done way better than current implementation too ;-)
21:04 RealBadAngel any code can be done better
21:05 RealBadAngel point is its easier to make better your own code
21:05 RealBadAngel because you understand it fully from the very start
21:06 sapier1 that's the difficulty with open source you have to take everything with you in order to get good code ... and of course learn from those who know better than yourself
21:08 celeron55 okay so
21:08 celeron55 i seriously want to get v5 running on 0.4
21:09 celeron55 having done singlenode, i'll probably do that
21:10 hmmmm you know what, i was wrong about 0.4.x being graphically slower
21:10 hmmmm i just tried 0.3.x and i get about the same FPS
21:10 hmmmm and oddly enough i have the same HORRIBLE HORRIBLE stuttering problems
21:10 celeron55 what hardware is that on
21:10 hmmmm but this is on my laptop, so i'd say the problem is probably because it's cpu bound
21:11 hmmmm amd sempron SI-42, geforce 8200g m
21:11 hmmmm 3gb of ram
21:12 Exio i still don't get how hmmmm computers work
21:12 RealBadAngel btw. https://www.youtube.com/watch?v=cjHNPzv9sGY
21:12 RealBadAngel this song we can freely use for minetest
21:12 RealBadAngel and others make by Skaven
21:13 celeron55 minecraft tends to stutter like a piece of smelly shit too on some hardware; for example on this i3 integrated GPU it's completely unplayable at least on windows
21:14 celeron55 minetest is better than minecraft on windows on this, and on linux minetest works pretty much flawlessly
21:15 celeron55 i think it happens at when new meshes are drawn the first time
21:15 hmmmm nah
21:15 hmmmm the stuttering i'm talking about happens all the time
21:15 hmmmm you can most notice it when you turn fast mode off, fly mode on, and just slowly go through the air
21:16 celeron55 well, that's the kind of stuttering i see - it happens all the time because a mesh is generated for each received mapblock
21:16 hmmmm and then at times there's a big stutter where it'd stop completely for about 2/3rds of a second it must seem like
21:16 hmmmm and these blocks are already loaded might i add
21:16 celeron55 well that's way worse than i've seen
21:16 VanessaE hmmmm: is preload visual turned on?
21:17 VanessaE actually ignore that, probably not related
21:17 VanessaE (only time I see those huge fps dropouts is when the engine stops to render a new block)
21:18 VanessaE (e.g. load a new texture that hasn't been used yet)
21:18 RealBadAngel folks, what do you think of makin this tune Main Menu score? https://www.youtube.com/watch?v=cjHNPzv9sGY ?
21:23 celeron55 this might be the least horrible music suggestion i have seen
21:24 VanessaE "least horrible"? you know, there are words on the other side of the center such as "best" :p-)
21:24 VanessaE :-)
21:24 hmmmm the negativity is just what makes celeron.... celeron.
21:24 RealBadAngel this is piece of music made by one of the best musicians for PC demoscene ever
21:25 RealBadAngel walking and living legend
21:25 hmmmm i can go for this
21:25 VanessaE I'll agree with it
21:25 VanessaE if we choose any tune, that one is good.
21:25 RealBadAngel google and youtube for Skaven/Fairlight
21:26 VanessaE (of course, RBA knows I am slightly biased ;) )
21:26 RealBadAngel Skaven has granted us rights to use any of his tracks that wasnt used in commercial projects
21:30 celeron55 i think i'm ok with trying, but there are questions
21:30 celeron55 where do we put the... track?
21:31 RealBadAngel hold on
21:31 celeron55 like, it's obvious it should be in windows build .zips (or is it?) but like otherwise
21:33 RealBadAngel celeron55, https://github.com/minetest/minetest/blob/master/src/game.cpp#L846
21:33 celeron55 yes and?
21:33 celeron55 does everyone rip it from youtube themselves
21:33 RealBadAngel ofc not
21:34 RealBadAngel i do have normalized oggs
21:34 celeron55 do we make a "minetest music" repo?
21:34 VanessaE just as long as I can turn the music off easily of course
21:34 RealBadAngel so add second slider
21:35 RealBadAngel and treat music as another instance
21:35 Exio http://www.futurecrew.com/skaven/index2.html
21:36 RealBadAngel Exio, Skaven is a demoscene coder like me, aged also 40
21:37 RealBadAngel we had a long talk bout past and what we are doing today
21:38 RealBadAngel when i asked him for the tracks he said no problem, just dont use those i sold to commercial gamas
21:38 Exio hehe
21:40 RealBadAngel celeron55, this piece of code is a good basis to create pack of default music scores with preset names
21:40 RealBadAngel distributed along with the game
21:42 celeron55 multiple tracks sucks to distribute though
21:42 RealBadAngel or overwritten in /sounds/all
21:42 PilzAdam I dont like this music
21:42 celeron55 the package size gets immediately multipled many times
21:43 RealBadAngel cmon that is not a problem nowadays
21:43 VanessaE storage and bandwidth are cheap
21:43 sapier1 bandwidth isn't cheap!!!
21:43 RealBadAngel i can dowload 900kbs
21:44 VanessaE I can pull down 2.5 MB/sec
21:44 RealBadAngel and it costs me like 15$ per month
21:44 celeron55 but sometimes you just happen to be behind a wlan that only pulls 80kB/s - then downloading some 15 megabyte packages isn't really something you want to do
21:44 VanessaE costs me about US $42/mo
21:44 celeron55 or multiple tens of MB
21:44 sapier1 and I'm at 200 may and I don't have any chance to get more because it's not available where I live and most likely won't be available anytime soon
21:45 VanessaE so make it a separate "music pack"
21:45 VanessaE as mintest_game is separate from the engine
21:45 Exio [18:43:32] storage and bandwidth are cheap
21:45 RealBadAngel celeron55, first of all we are makin a GAME
21:45 RealBadAngel not just a bunch of c++ code lines
21:46 ShadowNinja A separate package sounds like the best idea
21:46 Exio here, when my 200mb/mo ends it starts using like 1 USD / hour for a basic inet,
21:46 celeron55 good and well selected music (without some audiophile quality) is always worth the size
21:46 celeron55 but not some random tracks
21:46 Exio my 3g, the best connection what i can have, and really, if i can download less stuff, better
21:46 celeron55 (not implying anything)
21:46 PilzAdam why not tell people to get a music player?
21:47 sapier1 recently german telekom announced to limit "flatrates" to 75 gb ... most likely any other provider will follow soon ... seems time of cheap internet is over
21:47 RealBadAngel /giveme technic:music_plyaer 1
21:47 RealBadAngel :P
21:47 PilzAdam why do we need everything in Minetest?
21:48 PilzAdam they can run their music player in the background
21:48 RealBadAngel because we are not DEAF? ;)
21:48 PilzAdam and everyone picks the music they like
21:48 RealBadAngel PilzAdam, games, titles are coming with tracks
21:49 PilzAdam but most music I heard in FOSS games is horrible
21:49 RealBadAngel game without content, music is just..... naked?
21:49 celeron55 naked games are sometimes good
21:50 RealBadAngel good naked is nice chick
21:50 RealBadAngel but not a game :)
21:50 Exio i don't see the why not if it is optional at "some point"
21:50 celeron55 i do think minetest isn't really fitting to any model that somebody would say a good game is like
21:50 PilzAdam how is the progress of the web browser inside minetest?
21:51 PilzAdam it will be optional too!
21:51 Exio there is already code for doing the stuff, for what i understand
21:51 RealBadAngel dont mix the ideas Adam
21:52 RealBadAngel celeron55, btw streamin audio is also on its way, somebod else is doin this, seen in forums
21:53 celeron55 currently minetest looks a bit like doing that to technology what minecraft did to gameplay, and doing that to gameplay what minecraft did to technology 8)
21:53 ShadowNinja PilzAdam: hmm, merge the lua file download pull, convert html to formspecs and presto, basic web browser
21:53 sapier1 lol "convert html to formspecs"
21:53 celeron55 http://en.wikipedia.org/wiki/Presto_%28layout_engine%29
21:54 celeron55 so then we'll overtake that name as opera is dumping it :-----D
21:54 RealBadAngel celeron55, as long we going forward it doenst make a differnce, does it?
21:54 celeron55 RealBadAngel: how do you know you're going forward or not?
21:55 RealBadAngel makin your home nicer, with better stuff?
21:56 RealBadAngel and each time having fun of testing new piece from community fellows
21:56 celeron55 okay so maybe we need to add a game called "dollhouse" which will contain the music and such
21:56 RealBadAngel that is movin forward
21:57 PilzAdam my wishlist: minetest should be 1) a web server 2) a music player 3) a text editor 4) a web browser 5) a c++ IDE 6) a compiler
21:57 RealBadAngel lol
21:57 sapier1 you forgot the allknowing coffee making ai
21:57 celeron55 seriously, i'm having hard time seeing where we're going and why that is a good goal
21:58 celeron55 you can "go forward" for years and end up with something nobody has any use for
21:58 celeron55 s/we are/RBA is/
21:59 RealBadAngel 1) possible 2) already done) 3) ouch) 4) again?? 5) theres Z80 CPM with c++ 6) and compiler also
21:59 sapier maybe we shoud start a forum thread where everyone writes where she/he sees minetest in 1y 2y 5y?
21:59 celeron55 sapier: is that a joke? :D it'll be full of completely differing things
22:00 celeron55 and there has probably been one already
22:00 RealBadAngel like every wishlist
22:00 Exio PilzAdam: replace minetest with emacs, then :P
22:00 PilzAdam celeron55, what about this? https://github.com/minetest/minetest/issues?labels=bug&page=1&state=open
22:01 sapier :-) I expected this to be completely different, but maybe knowing in wich direction each of the core developers is working might explain some things ;-)
22:01 celeron55 but in all seriousness, minetest 0.4 is quite good at allowing people build things
22:02 celeron55 so i guess that has to be part of what will be aimed for
22:02 celeron55 as far as other things go, that is somewhat unclear
22:03 sapier imho we need to fixup locking within minetest in order to gain (more) benefit of multicore cpus
22:04 celeron55 but really, we should have a goal or philosophy or something to which compare everything against
22:04 sapier I fully support this
22:05 celeron55 i think that has formed to be "make a thing that kind of continues the path set by minecraft alpha, additionally allowing people to extend and modify it as easily as possible"
22:06 sapier good idea ... and how do we want to reach this?
22:06 PilzAdam by waiting for someone to code it?
22:07 PilzAdam -? +!
22:07 sapier I'm more interested in the philosophy minetest core wil follow
22:07 celeron55 it's interesting that roughly everyone who has played MC alpha knows what that means
22:07 celeron55 it's quite a strong thing
22:08 sapier maybe this is because everyone experienced mc alpha in a different way?
22:08 RealBadAngel celeron55, only imagination is the limit. thats the ultimate goal of a sanbox
22:09 PilzAdam RealBadAngel, no, you have to set yourself limits, otherwise its boring
22:09 celeron55 but really, i wonder whether we need more than that pseudo-quote
22:09 ShadowNinja hmmmm: I heard that you accidentally fixed black leaves on trees, if that is so what did you fix it in?
22:09 RealBadAngel settin no limits is even harder
22:10 RealBadAngel you have to imagine what other player would do
22:11 hmmmm ShadowNinja, i didn't, i thought i did but i was wrong
22:11 celeron55 then i wonder, how does this fit with music
22:11 PilzAdam celeron55, https://www.youtube.com/watch?v=bIOiV4d1SVI
22:11 celeron55 it doesn't really
22:11 Exio i don't see the *why not* if the code is already here and working
22:12 RealBadAngel same as with /textures/all
22:12 RealBadAngel new forums theme Sounds Packs
22:13 celeron55 PilzAdam: 9:32 is epic 8)
22:13 RealBadAngel and thats all.
22:14 RealBadAngel some may like defaults, some will insert metallica there or limp bizkit
22:15 RealBadAngel or (apaga satanas!) justin bieber
22:16 PilzAdam http://dev.minetest.net/Changelog#0.4.5_.E2.86.92_.3F <- the "Bug fixes" list is too small...
22:17 celeron55 i really think there are two kinds of people within this community: the ones who have come around to build stuff with the modding capabilities, and the ones who wish to have a "minecraft alpha with a twist"
22:18 celeron55 but i don't know how much they should be separated from each other
22:18 celeron55 their goals are imcompatible with each other to some extent
22:18 celeron55 incompatible*
22:19 Exio what about being of "both"?
22:20 celeron55 maybe you can answer that question
22:23 celeron55 bleh, i need to sleep
22:24 celeron55 however, that's the overall challenge here
22:24 celeron55 it tends to reflect in everything that happens here
22:28 celeron55 so that is an explanation to why i'm having hard time to decide on the music thing or the particular track!
22:28 celeron55 oh, then there is a third thing
22:30 celeron55 it is that if everything is made configurable and people are always expected to set up stuff in their own way, the community will always be about tinkerers who mainly develop the game for themselves and don't really aim for any praticularly finished end result
22:31 RealBadAngel celeron55, dont forget its sanbox
22:31 celeron55 i think we should leave room for a coherent game to emerge from all this that would be just that, a game, without being aimed to tinkerers like us, if somebody wishes to make one
22:31 RealBadAngel players are supposed to do so
22:32 hmmmm my feelings on the topic are basically, "here's a complete game, but look, it's very configurable, so if you don't like something, you can fix it yourself"
22:32 RealBadAngel abililty means it can fit anybody
22:33 Exio as i said, i don't see the point for a why not if it already coded (the "hard" work if we can say it in someway)
22:33 Exio just saying from what i have understand
22:33 PilzAdam my thoughts: "heres a complete game and its not configureable; go to the forum and install other games if you dont like it"
22:33 RealBadAngel even developers have different tastes, so what about players...
22:34 RealBadAngel PilzAdam, no way
22:34 PilzAdam (as for minetest_game, common, whatever that is)
22:34 RealBadAngel not for a sandbox one
22:34 RealBadAngel it is all abot the FREEDOM
22:35 celeron55 for one thing, i basically hate how little "brand" the individual games inside minetest are able to have
22:35 celeron55 but i don't see a good way of changing it, and in all cases it doesn't even make sense to change it
22:36 RealBadAngel celeron55, we will unite this thingies soon thx to portals (or whatever Jeija called them)
22:36 celeron55 ...eh what?
22:36 RealBadAngel once you are in game, connnected somwhere you will be able to go elsewhere
22:36 RealBadAngel to another world
22:37 RealBadAngel with different rules
22:37 celeron55 wtf, how is that related to anything i said
22:37 RealBadAngel "individuals" and "brands"
22:38 RealBadAngel now imagine switchin brands in no time
22:38 RealBadAngel just by walkin
22:39 celeron55 that is exactly the problem
22:39 hmmmm by "in no time", you mean "the time it takes to download all the media"?
22:39 hmmmm ugh we need the TCP fast
22:39 RealBadAngel yeah
22:40 hmmmm and no stupid "let's embed a webserver into minetest because we're too incompetent to get TCP working"
22:40 hmmmm i still can't get over that suggestion
22:40 hmmmm who came up with that idea again?
22:40 RealBadAngel i was sayin somethin like this but lets say i dont remember ;()
22:40 VanessaE hmmmm: other way around, you mean "embed a wget equivalent"
22:41 VanessaE and i don't remember whose idea it was
22:41 VanessaE but you're right, TCP is needed.
22:41 celeron55 it's only equally crappy as minetest having libcurl for downloading media from HTTP servers
22:41 hmmmm no, they wanted to run a webserver to host the textures and whatever
22:42 RealBadAngel ok ok, it was me. shoot me
22:42 VanessaE hmmmm: well it was my suggestion that if we have a wget equivalent, that we should have put a minimal webserver at the other end -- but I also assert that neither should be needed *at all*
22:42 VanessaE instead, it's wget at one end, and a third-party separate webserver at the other end
22:42 VanessaE which is fucking stupid
22:42 celeron55 eh, i have said the same thing sometimes too
22:43 celeron55 TCP is basically waiting to happen, but nobody has the balls to try 8)
22:43 hmmmm ..or they didn't work on it
22:43 VanessaE celeron55: well you kinda talked people out of it by saying how horrible your TCp implementation is :)
22:44 celeron55 it was a hack to try to get it quickly done as a start, it isn't anywhere near what a real implementation should be
22:44 hmmmm i know that if i were working on it, i would definitely get it to happen, but i'd have pause everything else i'm doing at the moment
22:44 hmmmm have to pause*
22:44 RealBadAngel there are no tcp libraries out there?
22:44 celeron55 yes there is, it's called posix sockets
22:45 celeron55 8)
22:45 celeron55 you don't need more than that unless you want more problems
22:45 hmmmm i wish i could make myself multithreaded
22:46 VanessaE while true; do launch.instance(hmmmmZ) & done
22:46 VanessaE shit;.
22:46 celeron55 hmmmm: what are you in the middle of now, then? didn't you just get ores done?
22:46 hmmmm instead, i think i waste a lot of time switching to do different tasks
22:47 hmmmm yeah, I was doing something on my own that I wanted to get done real soon
22:47 hmmmm i found an old minecraft classic map file of my first ever map and i wanted to ressurect it
22:47 celeron55 ->
22:47 hmmmm ->
22:48 hmmmm -->
22:48 VanessaE <--
22:48 hmmmm -^-
22:49 RealBadAngel lol
22:49 VanessaE heh
22:49 sapier >>>><<<<
22:50 RealBadAngel btw, when do we want 0.4.6 to happen?
22:50 hmmmm the plan was right after pilzadam adds the new ores
22:50 PilzAdam the new swords dont do any damage currently
22:50 VanessaE yesterday
22:50 PilzAdam and they wont be until the new damage system is done
22:51 RealBadAngel im holding techinic changes for it
22:51 VanessaE why do we need a new damage system now? everything worked fine before, if unideal
22:51 RealBadAngel i realized what c55 meant with rolling releases
22:51 hmmmm it was explained a couple days ago
22:51 PilzAdam if you want to do the fleshy groups, go ahead
22:51 PilzAdam I wont
22:51 hmmmm so basically we're waiting on celeron
22:52 PilzAdam VanessaE, they are marked as TODO, relatively easy to find
22:52 VanessaE youmissed my point
22:52 VanessaE I mean why do we need it *first*
22:52 VanessaE what happened to the idea of releasing something that at least works as well as it always did, and improving it later?
22:53 PilzAdam because I dont want to release swords that dont cause damage
22:53 VanessaE then don't add the swords.
22:53 hmmmm nnnnnnnnnnnoooooooooooooooooooo
22:54 RealBadAngel what for swords now if there are no mobs yet???
22:54 PilzAdam also the picks, axes and shovels dont have fleshy group defined yet
22:54 PilzAdam so leave them out too?
22:54 hmmmm PvP
22:54 VanessaE *facepalm*
22:54 PilzAdam swords are currently only good for PVP and digging leaves
22:54 VanessaE and that's all they can be used for anyway
22:55 VanessaE RBA is right - we have no build-in MOBs..what does it matter right now?
22:55 VanessaE built*
22:55 RealBadAngel so fix the stupid system that you have to turn aroun, lose focus to hit again
22:55 RealBadAngel without it, dont even mention swords again
22:56 PilzAdam someone can make the fleshy groups
22:59 sapier is anyone interested in reviewing the scriptapi fixes before I issue a pull request?
22:59 VanessaE can we just agree not to overcomplicate this?
23:00 VanessaE all I know is if I pick up an evil sword (gloopblocks), it takes maybe two to four strikes of that sword to frag a player or destroy a peaceful NPC
23:00 VanessaE where is the problem here?
23:03 PilzAdam the problem is that the current system is messed up and I dont want to add random numbers in there
23:03 PilzAdam so you basically have to test everything in-game
23:03 PilzAdam wich is a PITA
23:03 VanessaE then. don't. add. swords.
23:03 VanessaE stop trying to change how shit works
23:04 PilzAdam but all tools cause damage
23:04 VanessaE people already understand that an axe or a pick doesn't work as a weapon
23:04 PilzAdam (thx to Calinou)
23:04 VanessaE and they already understand how to use what *does* work as a weapon
23:06 PilzAdam can someone ask Calinou to do it when hes arround? he has done all the other fleshy groups too
23:07 ShadowNinja Can someone please pull this, it has even been rebased for easy merging: https://github.com/minetest/minetest/pull/421
23:07 PilzAdam hmmmm wanted to look at it IIRC
23:08 PilzAdam I tested it after rebasing and it works fine for me
23:08 hmmmm what? it was fine after the syntax got fixed and what not
23:08 hmmmm err code style
23:09 PilzAdam ok, Ill merge it then
23:12 RealBadAngel http://realbadangel.pl/technic_track8.ogg
23:17 RealBadAngel http://realbadangel.pl/technic_track5.ogg
23:26 Taoki I can't find out what's happening
23:33 Taoki PilzAdam: Seems to work in the Build game. minetest_game won't start any more though... says it can't load init.lua
23:34 Taoki And I figured out why it only worked for console... I was running the system package not the GIT. So it's a problem in latest hit
23:34 Taoki ***git
23:35 Taoki 01:33:01: ERROR[main]: Server: Failed to load and run /windows/D/Small games/Minetest/Minetest_GIT/bin/../games/minetest_game/mods/default/init.lua
23:35 Taoki 01:33:01: ERROR[main]: ModError: Failed to load and run /windows/D/Small games/Minetest/Minetest_GIT/bin/../games/minetest_game/mods/default/init.lua
23:35 Taoki And yes that is the correct path
23:36 Taoki Erm... init.lua is actually gone from there o.o
23:36 VanessaE because it's part of common now.
23:36 VanessaE https://github.com/minetest/common
23:37 Taoki ok. So minetest_game is gone, and it's called common now?
23:37 VanessaE no
23:37 VanessaE it just uses common
23:37 VanessaE if you use minetest_game, you need common also.
23:37 Taoki ok. Why that error then?
23:37 Taoki I got it now
23:37 Taoki So it should work
23:37 ShadowNinja Taoki: delete all empty mod folders in minetest_game, git doesn't remove them
23:39 Taoki Oh thanks, that fixed it
23:40 Taoki I don't understand the latest changes as well, but glad to have it working again
23:40 Taoki Like why minetest_game was split up like that
23:59 kaeza Taoki, because of the split to games/survival and games/build, having common eases updating (you don't have to update some common mod more than once)