IRC log for #minetest-dev, 2013-04-09

← Previous day | Channels | #minetest-dev index | Today | Next day → | Google Search | Plaintext

All times shown according to UTC.

Time	Nick	Message
00:42		dexter0 joined #minetest-dev
01:37	hmmmm	pilzadam, looks like it might be
01:37	hmmmm	isStaticAllowed wasn't already implemented by ServerActiveObject? :\
01:38	hmmmm	didn't realize that what I wanted could be done so smoothly...  i have to take a better look at this though
01:39	hmmmm	and this is only one part of it, we still need to set staticize for drops such as saplings and mined blocks to false
01:40	hmmmm	proller, why not run it with the whole thing in Debug?  also, there aren't any other divisions in generateCaves(), so i don't know how that's possible
01:42	salamanderrake	why did minetest start to get slower then minecraft, not nocking just wondering.
01:46	VanessaE	slower how?
02:07	VanessaE	Why does the code for growing saplings -> trees use the mapgen aliases instead of the real node names?
02:12	hmmmm	because that's the whole point of mapgen aliases
02:16	VanessaE	but growing a sapling into a tree isn't done at mapgen tine.
02:16	VanessaE	time*
02:29	VanessaE	am I right in that it is impossible to disable, delete, or otherwise remove default saplings from the game?
02:29	VanessaE	from a mod that is.
02:32	hmmmm	hmm boy you really hate the default trees don't you
02:32	VanessaE	it's not that I hate them.
02:33	VanessaE	it's that I wanted a single, central to replace them with something else.
02:33	hmmmm	no, you can remove it from the game by modifying the mapgen aliases for leaves, tree, and saplings to air
02:33	VanessaE	I already tried that, it doesn't actually work for saplings.
02:33	VanessaE	(for trees, leaves it works fine)
02:33	VanessaE	minetest.registered_items["sapling"] = nil
02:33	VanessaE	minetest.registered_aliases["sapling"] = nil
02:33	VanessaE	minetest.registered_aliases["mapgen_sapling"] = nil
02:33	VanessaE	minetest.registered_aliases["mapgen_sapling"] = nil
02:33	VanessaE	minetest.registered_items["default:sapling"] = nil
02:33	VanessaE	minetest.registered_aliases["default:sapling"] = nil
02:34	VanessaE	if that doesn't nuke them, nothing will.
02:34	hmmmm	so what's wrong with just letting the player not be able to get a regular sapling in the first place?
02:34	VanessaE	I tried that too.  Something elsewhere in the engine or game code is overriding my code.
02:35	hmmmm	if you aliased leaves to air, there's no way you'd be able to get saplings without /give
02:36	VanessaE	creative mode.
02:36	hmmmm	if it's a creative mode, are you right in banning regular trees from the user?
02:36	VanessaE	the whole point is to rename, redefine, and replace them with something that I think is a little better - but using the default node names for compatibility.
02:37	VanessaE	oh well.
02:42	hmmmm	erm anyway
02:42	hmmmm	what if we made trees grow gradually
02:43	hmmmm	this can be done by having a special growing tree node that drops regular wood when mined, but has an ABM attached to it that scans upward to find the current height of the tree
02:43	hmmmm	when it gets to a certain height, it stops growing
02:43	VanessaE	I'd have to say no on that
02:44	hmmmm	why, because it conflicts with your mod?
02:44	VanessaE	it precludes the idea of a tree that can branch out from the center trunk
02:44	hmmmm	you can do that if you'd like
02:44	hmmmm	the vertical height from the center is just a way of gauging its current stage of growth
02:45	hmmmm	besides, it's better than nothing
02:53		salamanderrake joined #minetest-dev
04:18	hmmmm	i just can't do this anymore
04:18	hmmmm	school, minetest, and the rest of life is like a full time job
04:19	hmmmm	if only i were multithreaded
04:19	hmmmm	basically minetest comes to a halt when i'm busy with other things, i really wish that wasn't the case
04:20	hmmmm	but i wonder if that's so bad... is minetest growing too fast, or not fast enough?
04:20	hmmmm	i feel like not fast enough, but that's only from my perspective... others can't seem to ever have a stable release from git
04:21	hmmmm	what do all of you guys think about the current pace of minetest development?
04:22	VanessaE	I think it's fine in fact.
04:22	VanessaE	If you need to slow down, do so
04:22	VanessaE	don't burn out.
04:23	hmmmm	no, i'm not burning out at all.... i just wish i could spend more time doing things that need to get done
04:23	VanessaE	ah
04:24	hmmmm	like the lua MapVoxelManipulator is something that everybody needs, like right now, but it's so far off because of all the other things that need to be done first
04:24	hmmmm	optimizing mobs is another thing that we need *right now* but we'll probably never get to
04:25	hmmmm	and none of this can happen until i get done with the myriad of other things on my TODO list and the endless flood of shit from school
04:25	hmmmm	i'm so lucky i don't have to study for anything, because then i'd have no time at all for real
04:26	VanessaE	and here I have tons of free time but I just can't get my head wrapped around C++
04:26	VanessaE	with I could "donate" some of that spare time :-)
04:27	hmmmm	it's really not that hard at all
04:27	VanessaE	you know, go off to sleep and you use my extra shares :)
04:28	hmmmm	and you'd probably make a good developer for minetest, too, if only because you're already familiar with the way things work
04:37	hmmmm	I think I have a solution to the whole Lua init/mapgen params mess; i would like to add another callback called on_mapgen_init() to be called once after the mapgens have been created, in the Server ctor, this is where you'd be able to register the mapgen-specific on_generate along with any mapgen-specific ores you may want to add
04:37	hmmmm	i'll do this sometime during the week
04:50	VanessaE	nice
05:20	VanessaE	here's an idea:  Stuff in the distance should not appear just in front of the fog.  Rather, it should fade in.  I think this could be accomplished by simply drawing the fog "closer"
05:20	VanessaE	(relative to the view distance)
05:21	VanessaE	or make the fog get thicker in the distance than it does now, something like that.
05:26	celeron55	07:19:53 <+hmmmm> basically minetest comes to a halt when i'm busy with other things, i really  wish that wasn't the case
05:26	celeron55	i don't think so
05:27		BackupCoder joined #minetest-dev
05:27	hmmmm	explain why there have not been any commits in the that i either didn't do myself or people waited for me to give it the okay to do it
05:27	hmmmm	in the past i don't even know how long*
05:28	celeron55	minetest is about more than just commits to core 8)
05:29	hmmmm	people can develop all they want, but it needs to get committed... and reviewed, and tested, and all of the things that come along with it
05:29	celeron55	also, people will like to use you as the okay-giver because you're currently using the most time for developing core; they'll ask someone else when you're not around
05:29	hmmmm	it's practically nothing in comparison to the effort necessary to actually write the code, but nevertheless it takes time
05:30	hmmmm	when people want things committed, whatever i'm working on comes to a halt too.  i counted only 2.5 days worth of quality time to work on mapgen v7
05:32	hmmmm	it's good that i am looking over everything of sapiers' memory leak fixes though, because i caught a couple things that would've caused a lot of problems had they been blindly merged
05:32	hmmmm	i don't know, maybe it's fine this way and i'm just frustrated because i can't do more of what i like to do
05:32	celeron55	i suggest just not caring about other things when you want to work on it - at some point minetest can cross the line of one person not being able to handle everything that goes upstream
05:34	hmmmm	ha ha... you mean like right now?
05:34	hmmmm	the pull requests are backing up
05:35	celeron55	well, i have tried to choose all core developers from such people who are generally wise enough to not merge things they aren't able to check (with their knowledge or time)
05:35	hmmmm	if pilzadam weren't around it'd probably be up to 100 by now
05:38		emptty joined #minetest-dev
05:39	celeron55	churning through pull requests is the kind of thing most people won't do unless they're paid to do it
05:42	celeron55	because... well, you have so much responsibility and need so much knowledge that you could use much more creatively elsewhere
05:42	celeron55	and at the same time you need to tell most people they suck at coding
05:43	celeron55	really the only thing that would scale here would be something like linux does
05:44	celeron55	but as it doesn't seem to emerge out of nothing even in a very tiny extent, i guess it's impossible
05:44	hmmmm	doesn't linux have 3 people at the top deciding what gets merged?
05:44	hmmmm	that's really bad, is it not?
05:44	celeron55	linux kind of has a "tree of trust"
05:46	celeron55	the 3 people at the top don't really need to check much anything; they just get large bunches of patches from maybe 3 people each, who in order get similarly more from other people; each person manages some subsystem and knows the quirks of the code and the people who send patches to them
05:46	hmmmm	well, i trust pilzadam completely to do all minetest_game related things and most of the core things
05:46	hmmmm	8)
05:46	celeron55	it will scale infinitely as long as new people "attach" to the lower end of the tree
05:46	celeron55	we don't have even a single layer of such
05:47	hmmmm	...perhaps we should start making something like that
05:47	celeron55	it could be worth a try, but i'm somewhat pessimistic of the possible results
05:48	hmmmm	why, not enough qualified people?
05:48	celeron55	but basically how to start is pointing core devs to certain parts of the code, and telling all new contributors to note their pull requests to the respective people
05:48	celeron55	and stating that the goal is to scale like linux
05:50	celeron55	i once (or twice) suggested this before, but as the idea took zero traction, i just threw it away
05:50	celeron55	now as there are two liking it, it might work 8)
05:52	celeron55	i do believe that you can get a lot more motivation from people if you give them actual responsibilities, of the size that they are able to manage - but the starting point to that needs to be that there are capable enough people
05:52	hmmmm	right
05:52	hmmmm	who here is capable..
05:52	hmmmm	and more importantly, at what
05:53	hmmmm	we already know the answers to that though
05:53	celeron55	actually it doesn't even require much skills, other than knowing what one *doesn't* know, so one can ask those who know and learn at the process
05:53	celeron55	that skill seems to be a lot asked these days though :-D
05:53	celeron55	"I KNOW C++ WHAT DOES i++ DO?"
05:54	hmmmm	indeed, a lot of people seem to think that they just pick up C++ and then they can start developing
05:55	hmmmm	well
05:55	hmmmm	that's what i did
05:56	hmmmm	but that's different
05:56	hmmmm	i already knew how to code
05:57	hmmmm	yes, that's the difference... there are those who know how to code, and those who know a language
05:58	hmmmm	what makes minetest so attractive to beginning developers, who lack the former?
05:58	VanessaE	Lua.
05:59	VanessaE	one easy language somehow translates to others being equally easy.
05:59	VanessaE	(even if it doesn't)
05:59	hmmmm	no, it's not that, i see the kind of people celeron is talking about on the forums and such
06:00	hmmmm	the subset of people who contribute right now, i'd say, know both how to code and the details enough to be effective (if they weren't effective coders, we wouldn'tve accepted their code in the first place)
06:00	celeron55	they're young
06:01	celeron55	people will know what they don't know much better when they get 18 or 20 or so
06:01		ssieb joined #minetest-dev
06:02	hmmmm	heh, it's probably good that I don't have as much of a background with C++, so I can be the voice of sanity when people want to go overboard with messy C++ features
06:02	hmmmm	if i had to do it all over again though, i'd do minetest in C
06:03	hmmmm	there's just something about C++ that makes it feel dirty
06:05	celeron55	it's dirtily powerful too though
06:05	ssieb	I'm trying to package minetest, but there's a problem.  The common mods don't get put into common, so the server doesn't start.  Obviously, I moved them over, but shouldn't they get installed properly?
06:05	hmmmm	at the same time, i guess it's nice that i am forced to work with C++, and i gained another skill in the process
06:05	celeron55	like, in terms of if you know what you're doing, you're going to get much more of everything out of your code than in C
06:06	celeron55	ssieb: the CMakeLists.txt of 0.4.6 lacks the commands to do that; that was added slightly after
06:06	hmmmm	i know what you mean, everything takes longer when you're doing minetest-related code in C, but it isn't something that can't be solved with well-used macros
06:06	ssieb	I have master and it still doesn't seem to do it
06:07	celeron55	then i have no idea 8)
06:07	ssieb	the common directory isn't created anywhere
06:07	ssieb	it works for you?
06:07	* hmmmm	hasn't updated minetest_game since 0.4.4
06:07	hmmmm	no idea here
06:07	ssieb	oh :-)
06:07	hmmmm	the common directory was celeron's whole idea though
06:08	hmmmm	he *should* know
06:08	celeron55	but it was mostly implemented by other people!
06:08	celeron55	:D
06:08	celeron55	ssieb: you didn't really explain precisely enough for me to know what you are actually even doing
06:09	ssieb	following the compiling instructions
06:09	ssieb	I have master of minetest, I get master of minetest_game, put it in games, then run the compile and install
06:09	ssieb	where is the games/common directory supposed to come from?
06:09	celeron55	you need https://github.com/minetest/common too
06:10	ssieb	oh...  something to add to the compiling instructions?
06:10	celeron55	where are these outdated compiling instructions you are talking about?
06:10	ssieb	http://dev.minetest.net/Compiling_Minetest
06:10	celeron55	the amount of repositories sucks though, i think we should do something about that
06:11	celeron55	maybe an "integration repo" where everything in the main distribution would be thrown in
06:13	celeron55	well, added that one quickly there
06:13	celeron55	you might want to get the survival or build game too, but they aren't really necessary at the moment
06:15	ssieb	oh, those are separate repositories too?
06:15	* ssieb	checks the list
06:16	celeron55	yeah, welcome to git clone git clone git clone git clone git clone
06:16	ssieb	I see that part of the moreores mod has been incorporated, how will this interact in an existing world?
06:17	ssieb	e.g. there are two different recipes for bronze
06:17	ssieb	but tin wasn't included...
06:17	ssieb	so I can't remove the moreores mod
06:18	celeron55	i believe if you download the most recent moreores, it should handle that well enough in... some way
06:18	ssieb	oh good, ok, I'll try that
06:18	* ssieb	didn't think of that...
06:20	ssieb	yes, I see that was done.
06:29	hmmmm	the entire moreores wasn't added, we only picked the subset of features we all agreed on
06:29	hmmmm	those ores were aliased to moreores for compatibility
06:29	ssieb	I realize that, that's why I was concerned about the interaction
06:29	hmmmm	if you actually want moreores, you still need to download moreores
06:30	ssieb	but I see that the moreores author has modified it to handle that
06:31	* ssieb	is currently building rpm packages on several machines...
06:31	hmmmm	we work with popular mod authors if we add or do something that is relevant to their mod
06:32	ssieb	I find the server gets rather laggy when (I assume) there is flowing liquid nearby
06:32	* ssieb	hopes to get a chance to poke a debugger at it soon...
06:33	hmmmm	water flowing shouldn't be laggy under normal conditions
06:35	ssieb	something is making it really laggy sometimes and that's the only thing I can think of
06:35	ssieb	I turned off the cheat checks and that improved things a lot, but not entirely
06:36	hmmmm	what kind of hardware are you running minetest on?
06:36	ssieb	AMD Athlon(tm) 64 X2
06:36	celeron55	and are you running some laggy mod?
06:37	ssieb	is there a list of those? :-)
06:37	hmmmm	mobs
06:37	ssieb	no
06:37	celeron55	map generation mods
06:37	hmmmm	yeah those can and will hang up the server
06:37	celeron55	(which aren't laggy after the world has generated though)
06:38	ssieb	only moreores that I know of for map generation
06:38	hmmmm	hmmmm, it's possible that could be it
06:38	hmmmm	i hope not though
06:38	ssieb	and it's laggy when (I think) there shouldn't be any generation
06:39	VanessaE	fwiw, there was a time not long ago where the water pyramid on redcrab's server would cause a pretty significant performance loss just by being in view...
06:39	VanessaE	(and it was just wool and flowing water)
06:40	ssieb	I'll try with the latest build and see what happens
06:41	ssieb	the kids are really enjoying it, but they are getting a little frustrated with the blocks popping in and out while they're digging...
06:41	hmmmm	i've lived through that for a while and i know how frusturating that is
06:42	hmmmm	perhaps you can increase the server step interval to something less demanding
06:42	ssieb	I'll try that if it's still a problem
06:43	hmmmm	:(
06:43	hmmmm	i do all this optimization and in the long run it still doesn't help much
06:48		emptty joined #minetest-dev
06:52	ssieb	is there any plan to allow compiled mods?  or is that something you can do with lua without needing specific support in the server?
06:52	VanessaE	ssieb: LuaJIT support just made it into upstream in fact.
06:53	ssieb	upstream?
06:53	VanessaE	that's about as good an improvement as one can get for now, barring further improvements to the engine (or to LuaJIT itseld)
06:53	VanessaE	upstream = the Minetest source code repository
06:53	ssieb	I noticed there was some debate about whether that was actually faster or not
06:53	ssieb	I'm not familiar with lua so I was wondering if there was some way to include binary modules like python can
07:09	celeron55	the only way supported is making a patch to the core that implements something to call from mods
07:10	celeron55	we don't really see much benefit from supporting a binary interface
07:12	celeron55	well, if you insist on writing some kind of binary modules, lua supports such like python and you can wrap them to the existing interface with lua
07:13	celeron55	there exists a minetest chat->irc wrapper written somehow like that
07:14		emptty joined #minetest-dev
07:15	ssieb	oh, right, I did see that one
07:28	ssieb	well, my rpms are built and the server works properly, so it's time to sleep :-)
07:35	emptty	ssieb: can you please point me to your repo?
07:35	emptty	I'm currently updating the debian packaging, and I was wondering if you could have some interesting patches ^^
07:46	VanessaE	or use my build :)
07:59		BlockMen joined #minetest-dev
08:00	BlockMen	celeron55, you should rename the wiki to "Minetest", because now the pagetitels are e.g. "Minetest Wiki - Download" ;)
08:01	BlockMen	and when i tried to register, it did not send the email with pw, also not when i clicked "send new password". could you chek that too?
08:04		emptty joined #minetest-dev
08:06		BlockMen left #minetest-dev
08:10	celeron55	BlockMen: renamed; and about e-mail sending: i tested it yesterday and it did work
08:11	celeron55	maybe you have some overly jealous spam filter that doesn't like the way it sends e-mail via a regular domestic ISP's server
08:16	VanessaE	celeron55: "overzealous"
08:16	VanessaE	bbl
08:17	celeron55	ehm... well whatever, some word like that 8)
08:22		jin_xi joined #minetest-dev
09:32		ImQ009 joined #minetest-dev
10:06		Jordach joined #minetest-dev
10:06		Jordach joined #minetest-dev
10:24		proller joined #minetest-dev
10:49		JesseH joined #minetest-dev
10:52		Calinou joined #minetest-dev
11:14		ImQ009 joined #minetest-dev
11:35		Jordach joined #minetest-dev
11:35		Jordach joined #minetest-dev
11:43		emptty joined #minetest-dev
12:16		PilzAdam joined #minetest-dev
12:20	* VanessaE	is back.
12:59	Taoki	So after two days of work, I can reveal my (hopefully pleasant) surprise for Minetest: http://forum.minetest.net/viewtopic.php?pid=82728
13:00	Taoki	This is hopefully a mod that will allow many new exciting things, such as sharing buildings via IRC or forum and creating villages or cities in mapgen easily
13:00	Taoki	Still a lot to do though
13:06	celeron55	i don't see that as anything more than what worldedit already does
13:07	celeron55	you could've used worldedit's format and went directly for the generation thing
13:07	celeron55	(and what it does, worldedit does better - WE saves node metadata too)
13:08	PilzAdam	object collission is broken
13:08		jin_xi joined #minetest-dev
13:09	PilzAdam	the rat in my mobs mob gets constantly pushed upwards
13:09	PilzAdam	collisionbox = {-0.25, -0.175, -0.25, 0.25, 0.1, 0.25},
13:09		blue42u joined #minetest-dev
13:10		hmmmm joined #minetest-dev
13:13	jin_xi	a propos generating structures: here is an example of structures made with turtle graphics. i think that could be a flexible way of adding all sorts of stuff to biomes: http://i.imgur.com/qhWjgHq.png
13:17	Taoki	celeron55: Not sure how worldedit works, but likely not the same way as this mod
13:17	Taoki	Wanted to go for a simple implementation where you use a core and 3 markers to mark an area you work with, and allow working with text files. Then adding that to the map I/O
13:19	Taoki	This mod is also meant for importing / exporting a single structure per file to work with, and possibly allowing multiple folders of structures
13:21	Jordach	we should have a candy land biome
13:21	blue42u	We SHOULD have Lua biomes.
13:21	Taoki	Jordach: With mapgen v7 we can
13:21	Taoki	blue42u: ^
13:22	blue42u	YAY!
13:24	blue42u	Question: Should all gui-related lua command strings use the same syntax?
13:27	blue42u	i.e. Formspec looks like this: "list[current_player;main;0,5;8,4;]". Should other gui things use a similar syntax?
13:33	blue42u	Ok, more specific. I'm trying to add Lua hud to th game, but I'm not sure whether to change the syntax. Currently, its "inv~main~0~0~1~5", or somthing like that. Should I change it?
13:44		ImQ009 joined #minetest-dev
13:46		jojoa1997|Tablet joined #minetest-dev
13:53		blue42u left #minetest-dev
13:55		nyuszika7h joined #minetest-dev
13:57		JesseH left #minetest-dev
14:04		nyuszika7h joined #minetest-dev
14:36		BlockMen joined #minetest-dev
14:38		rubenwardy joined #minetest-dev
14:38	BlockMen	celeron55, i have disabled my spam filter and tried again, but i still get no mail. could you set a temp-pw that i can check mail adress and set pw manual?
14:45		serengeor joined #minetest-dev
14:48		emptty1 joined #minetest-dev
14:54		troller joined #minetest-dev
15:10		proller joined #minetest-dev
15:19		emptty joined #minetest-dev
15:23		Calinou joined #minetest-dev
15:42	hmmmm	oh yes
15:42	hmmmm	blue42u is around
15:42	hmmmm	too bad i don't have any time for minetest this week
15:42	hmmmm	it'd be great to get the bar api in
15:43		jin_xi joined #minetest-dev
15:43	Calinou	yells at paramat for not coding correctly
15:43	Calinou	wants the bar API to be in
15:44	hmmmm	like i said, i'd be basing the work off of his, not entirely using it
15:44	hmmmm	did i yell at paramat?  who is paramat btw
15:44	hmmmm	i see him on the forums but never irc
15:44	Calinou	the noise guy
15:45	hmmmm	prestidigitator?
15:45	hmmmm	he's not
15:45	hmmmm	they work together though
15:45	VanessaE	Calinou: nick recognition FAIL :)
15:45	Calinou	let's just call him the noise guy
15:46	VanessaE	paramat's the guy doing landup, canals, floatlands, etc.
15:46	hmmmm	from the code of blue42u i've seen, it seems quite reasonable
15:46	hmmmm	anyway we need more of a consensus on how the bar api should work, like should the server really have complete control over arbitrary drawing of elements on the client's UI and such
15:47	VanessaE	definitely not
15:47	hmmmm	that's what his lua hud api is able to do
15:47	hmmmm	a big detail in the bar api is to restrict where things are drawn
15:47	Calinou	VanessaE: server can already rickroll clients
15:47	VanessaE	it should be something where the mod can say "ok, I want a bar that measures X points with Z number of images, placed in bar slot Y", and the engine decides where slot Y is.
15:48	VanessaE	s/engine/client
15:48	VanessaE	/
15:48	hmmmm	that's sound though, and it can be disabled
15:49	VanessaE	and if slot Y is taken by a mod, it gets moved to slot Y+1 or something.
15:49	hmmmm	and it's not entirely about protecting clients against malicious servers, it's just making things neat and orderly
15:49	Calinou	but people with sound can still be rickrolled
15:49	Calinou	ah
15:49	hmmmm	indeed
15:49	hmmmm	what vanessa just said
15:49	VanessaE	limit the number of available slots to, oh, 6 including the default health bar should do it I think
15:49	hmmmm	the thing is, you can't rely on a mod knowing where to put the bars
15:49		BlockMen joined #minetest-dev
15:50	VanessaE	exactly, otherwise you have mods sticking them all over the damn screen, or over top of one another
15:50	celeron55	hmmmm: i think it should be very free
15:50	VanessaE	or partially off-screen, etc.
15:50	celeron55	hmmmm: we have no reason to restrict anything that is shown on the screen
15:50	VanessaE	maybe it can use the formspec style of positioning stuff
15:50		iqualfragile joined #minetest-dev
15:51	celeron55	hmmmm: if you restrict something, people will just complain about it; if you don't restrict, people are free to choose mods that behave sanely
15:51	Calinou	^ I agree
15:51	celeron55	instead of restrictions, we should provide means for mods to behave sanely
15:51	Calinou	can the bar API draw text anywhere on the client's screen? would be very useful too
15:51	hmmmm	yeah, i did expect people to complain
15:51	Calinou	(and it'd be even better if you could change text opacity)
15:52	hmmmm	the other thing is that you could provide placement hints
15:52	hmmmm	but
15:52	celeron55	for example, (i haven't read any of the pull request) there should be a way to stick elements to offsets from the sides and corners of the screen without guessing resolutions or anything
15:52	hmmmm	blue42u's lua hud api is more than bars and numbers
15:52	ShadowNinja	perhaps the option for manual placement or automatic?
15:54	celeron55	really the modern way to go about this would be to overlay webkit on top of the 3D renderings :P
15:54	hmmmm	the original conception of the bar api was to not only provide a way of stupidly drawing things onto a client's hud, but a way to internally maintain values that are set and the client responds to it by drawing the hud api on its own, not just modifying the hud string
15:54	celeron55	many recent games do that
15:54	ShadowNinja	celeron55: the new website design, will everything be transfered from your domain when all of the pages are made?
15:54	hmmmm	shadowninja, what's wrong with hints?
15:55	celeron55	ShadowNinja: transferred from my domain?
15:55	ShadowNinja	what do you mean be hints?
15:55	hmmmm	how about you provide the rect where it's drawn, but there's also an api to get the positioning of where the next bar should be
15:55	ShadowNinja	from c55.me
15:55	ShadowNinja	ah, yes, that sounds good
15:56	VanessaE	that I could agree with.
15:56	celeron55	ShadowNinja: well, www.minetest.net is currently the same machine as c55.me, so of course
15:56	VanessaE	as long as there's some way to know that a bar will always end up in a given position relative to the existing health bar or whatever
15:57	celeron55	i'll be looking for hosting of this elsewhere eventually though
15:57	ShadowNinja	ok, can I have the privs to edit it then?
15:57	VanessaE	hmmmm: what about a concept of vertical bars at the sides of the screen?  widescreen users might enjoy that
15:57	VanessaE	(I have no use for such, myself)
15:57	hmmmm	perhaps
15:57	celeron55	ShadowNinja: i think you do have now - but don't do that yet, there are some changes to be made still
15:57	hmmmm	it's all possible at this point
15:57	Jordach	VanessaE, i quite like the idea of vertical bars
15:58	ShadowNinja	ok, just tell me when it is ready
16:07	ShadowNinja	hmmmm: how long did default register_on_generated take befofe LuaJIT?
16:14	hmmmm	not much, because it didn't do very much after i moved ores out of there
16:14	hmmmm	the point of luajit is to speed up lua in general so it doesn't feel like a laggy piece of shit as much
16:15	hmmmm	basically everything is smoother now
16:16		dexter0 joined #minetest-dev
16:18		emptty joined #minetest-dev
16:35		rubenwardy joined #minetest-dev
16:42		BlockMen_ joined #minetest-dev
16:55		BlockMen joined #minetest-dev
17:16		BlockMen joined #minetest-dev
17:18		ssieb joined #minetest-dev
17:24		blue42u joined #minetest-dev
17:24	blue42u	So, what needs to be done to fix my hud api?
17:25	Exio	hmmmm: ^
17:25		Taoki joined #minetest-dev
17:28	ShadowNinja	blue42u: It was decided that mods should have a way to get a unoccupied position for a bar
17:30	blue42u	Does the syntax need to be changed?
17:30		Calinou joined #minetest-dev
17:31	ShadowNinja	I don't think so
17:31	ShadowNinja	better ask hmmmm
17:33	blue42u	Well, hmmmm?
17:35	celeron55	someone implement this tool in minetest: http://www.youtube.com/watch?feature=player_detailpage&amp;v=srj0yopOJ_Y#t=354s
17:37	VanessaE	celeron55: technic mining laser
17:37	VanessaE	MK3 can do something like that
17:37	VanessaE	but not as fancy
17:38	hmmmm	"fix"?
17:39	blue42u	Well, improve.
17:39	hmmmm	it's not really "fix" as it is "change" to fit with what we wanted
17:39	hmmmm	you see, we had the idea a while back to have a "bar api"
17:39	hmmmm	which does almost what your lua hud does
17:39	hmmmm	http://dev.minetest.net/TODO#Bar_API
17:39	blue42u	So... what do I need to do?
17:40	hmmmm	humm
17:40	hmmmm	i didn't see exactly how your code works for that, but from what i understand, you get a string that describes the HUD, and when health changes or something, you remove the bar and add the new one?
17:40	Exio	Set bar value (bar identifier, new value). Happens on a Lua defined-event (someone gets hit and takes damage, or w/e).
17:40	hmmmm	yeah?
17:40	Exio	(for what i see)
17:41	Exio	uh-hu, lag
17:41	hmmmm	but anyway
17:41	hmmmm	the big difference between the bar api and the lua hud is that the bar api maintains a counter for each of these bar values
17:41	hmmmm	both on the client and on the server
17:41	blue42u	Ah.
17:41	hmmmm	so the server wouldn't send a new bar to be created
17:41	hmmmm	it'd just update the value of an existing bar, given a bar ID
17:42	hmmmm	also
17:42	hmmmm	it was our idea to restrict where bars can be placed, but what will likely happen is that people would complain about the restrictions
17:42	hmmmm	but there can't be chaos on the client screen either
17:43	hmmmm	so what would be good is to have a get_next_bar_position() API that queries the lua hud subsystem for where it should place the next bar
17:43	hmmmm	but also this is client-dependent based on resolution and all that stuff
17:43	hmmmm	which complicates things
17:44	blue42u	So, if bar "data" == amount, and text "data" == text, and image "data" == image name, and inv "data" == list name?
17:44	Exio	get_next_bar_?? how would the server know the client's resolution? (what can change in any moment, if the client uses a window and resizes the screen)
17:44	hmmmm	erm, i don't mean to implement it like that
17:44	blue42u	Which is why I used porportional coords.
17:44	hmmmm	exio, the position would be proportional
17:44	hmmmm	yeah
17:44	Exio	ah, perfect
17:45	hmmmm	but the client would maintain a data structure with the style of the bar, bar position, etc.
17:45	hmmmm	your code right now is dirt simple
17:45	hmmmm	that's a good thing
17:45	Exio	i didn't know that, sorry then :P
17:45	hmmmm	but it's also not as flexible as it ought to be
17:45	hmmmm	(not as flexible as i had planned either)
17:45	blue42u	So... how flexible?
17:46	hmmmm	have the client and server maintain copies of a vector containing Bar objects, each of which have a texture name, texture pointer, value, position on screen, name string, etc.
17:47	hmmmm	keep in mind this what i'm thinking of off the top of my head
17:47	blue42u	"name string"? Or "id string"?
17:47	hmmmm	name string
17:47	hmmmm	the ID would be numerical
17:47	blue42u	Well, I always use id strings. :)
17:47	hmmmm	yuck
17:48	blue42u	When I can.
17:48	blue42u	My first language was Smalltalk. Can you guess?
17:48	hmmmm	not really, from what i've seen of your code, it's quite sane
17:49	hmmmm	just make sure you read this over http://dev.minetest.net/Code_style_guidelines
17:49	hmmmm	and they're guidelines, not absolute rules... this is just to preserve neatness and whatnot
17:50	blue42u	I've been coding Python for about 5 months, then working on robotics, then...
17:50	blue42u	I'm multi-lingual.
17:50	blue42u	I learned c++ in a week.
17:51	blue42u	See my anxiety?
17:51	hmmmm	erm
17:51	hmmmm	what matters most is that you know how to _code_
17:51	hmmmm	having prior experience with C/++ helps though
17:51	blue42u	yeah.
17:52	hmmmm	if your code's good, it'll be accepted, no matter what your experience is
17:52	blue42u	Anyway. "texture pointer"? We use strings to do that!
17:52	blue42u	== "texture name"
17:53	hmmmm	from what I recall, getTextureByName() locks a mutex evverrrrry single time it's called
17:53	blue42u	Really? I forgot this was multithreaded.
17:54	hmmmm	oh yes
17:54	VanessaE	hmmmm: why the hell would one use a lock on ...an image read routine?
17:54	blue42u	And, would "name string" be used for identification?
17:54	hmmmm	i'm not overly worried about it, but it's in poor taste and completely unnecessary inefficiency
17:55	hmmmm	no, the ID would be (this is for bars though, remember?)
17:55	hmmmm	your lua HUD api does more than bars from what i recall
17:55	hmmmm	vanessae, i'm not sure, i should probably check
17:55	hmmmm	*opens minetest*
17:56	VanessaE	heh, I meant generically.
17:56	blue42u	True. But 1) a numarical id could give a conflict and 2) the names arnt displayed anyway.
17:57	hmmmm	i was hoping we could eventually have a screen that'd show all the currently used bars
17:57	hmmmm	along with the names
17:57	hmmmm	if we were to add the name for the bar later, that'd make us need to change the protocol version again
17:57	blue42u	Now, THAT would be cool.
17:57	blue42u	Ill start writing my todo list again...
17:59	hmmmm	hmm, i guess the MeshMaking thread uses the atlas?
17:59	hmmmm	I don't know but it doesn't seem right now that you point it out
17:59	hmmmm	we should investigate further and remove this lock if it's unnecessary
18:01	hmmmm	and blue42u, nevermind about the texture pointer I guess... just store the texture ID so it doesn't lock *twice* at least
18:01	hmmmm	we'll fix that later if we do end up removing the locks
18:01	blue42u	k!
18:04	hmmmm	yeah, that's a general rule of thumb, especially for us.. if there's a more efficient way of doing something, and it doesn't hurt anything else by doing that, just do it
18:04	blue42u	Well, I tried.
18:05		serengeor joined #minetest-dev
18:05	blue42u	Does the bar data NEED to be known by the server?
18:05	hmmmm	it needs to be known by the mod at least
18:05	blue42u	Well, that would be the modders job, right?
18:05	hmmmm	I guess
18:05	PilzAdam	yes, there should be a function player:get_hud()
18:06	hmmmm	wait, why?
18:07		emptty joined #minetest-dev
18:08		sapier joined #minetest-dev
18:08	PilzAdam	mods should be able to restructure the whole thing, if there is no get_hud() then other mods' stuff will be lost
18:08	blue42u	It could just remember it itself.
18:09	blue42u	An extra var somewhere...
18:09	PilzAdam	sapier, line_of_site() and object <-> object collision is broken
18:09	sapier	remembering itself adds problem of recursion depth
18:09	sapier	how is line of sight broken?
18:09	sapier	los doesn't honor glass this is intended
18:09	PilzAdam	*line_of_sight()
18:10	BlockMen	could it be that costumizeable that a mod can change the complete hud and add new bars (like health bar) too?
18:10	celeron55	without reading your previous discussion, i think it could work so that when a mod adds something to it, it also always assigns a name for what it adds; then the parts could be removed or updated by name if needed
18:10	celeron55	and mods don't need to touch other mod's stuff
18:10	blue42u	== id
18:10	celeron55	so, it wouldn't even expose anything to get everything, or remove everything, or replace everything
18:11		jin_xi joined #minetest-dev
18:11	blue42u	BlockMen: Thats what we're talking about.
18:11	blue42u	Explain, celeron55.
18:11	hmmmm	actually there seems to be the opposite sentiment within the community
18:12	hmmmm	lots of people want to be able to interfere with other mods
18:12	hmmmm	so they can override things (which is a horrible policy to begin with)
18:12	sapier	Pilzadam can you tell me what s broken with object object collision?
18:12	blue42u	Give me the decition on a comment of the pull, I have to go.
18:12	Jordach	sapier, for some reason the rats in simple mobs literally fly away
18:12	hmmmm	line_of_sight doesn't return the right value or something
18:12	VanessaE	if I couldn't "interfere" with other mods, I couldn't redefine apples so that they fall.
18:12		blue42u left #minetest-dev
18:12	PilzAdam	sapier, things with little collisionboxes (like the rat in simple mobs) get constantly pushed upwards
18:13	hmmmm	vanessa, not you, others
18:13	PilzAdam	sapier, and line_of_sight() returns the step param
18:14	Jordach	quick question: why are caves, trees and dungeons disabled by default?
18:14	sapier	I'll fix line of sight ... how can small objects be pushed up?
18:14	hmmmm	caves and trees are disabled by default?
18:14	hmmmm	what world do you live in
18:14	Jordach	apparently its in the stable
18:14	Jordach	i have to append minetest.conf with mg_flags = trees, caves
18:15	Jordach	before they appear again
18:15	hmmmm	and?
18:15	Jordach	they return
18:15	hmmmm	did you read the default value for mg_flags?
18:15	Jordach	hm?
18:15	PilzAdam	sapier, I doesnt happen if I revert object <-> object collision commit, so you have broken it
18:15	celeron55	of course "mg_flags = " will remove them
18:15	celeron55	it's a setting value, not some "append these to the mapgen" thing
18:15	hmmmm	ugh
18:16	hmmmm	the faster i can get a UI for all of the mapgen things, the faster people will stop with not knowing what they're doing
18:16	hmmmm	i guess that's what i'm working on after mapgen v7
18:16	sapier	you seam to be very proud ;-) ... no matter I'll investigate and have a look if it's really an error
18:16	Jordach	celeron55, it's apparent that 0.4.5 worlds miss the mg_flags parameters
18:16	Jordach	same for stable builds as 0.4.6
18:17	celeron55	to mix up this discussion even more, this is now deemed to go live after it has all the necessary content: http://c55.me/minetest4/doku.php
18:17	celeron55	it uses this dokuwiki template, which can be polished if you find small errors: https://github.com/celeron55/minetest.net_dokutemplate
18:18	celeron55	if you don't have other use for your time, consider registering there and asking me privileges so you can fill in content
18:18	Jordach	that's a really nice template
18:18	ShadowNinja	celeron55: can I add content now?
18:18	celeron55	ShadowNinja: yes
18:18	celeron55	i belive you should have the privileges for that
18:18	hmmmm	the development dropdown menu has overlapping text
18:18	Jordach	celeron55, signed up as Jordach
18:18	celeron55	Jordach: did the password e-mail work? i have gotten a complaint about it
18:19	Jordach	celeron55, noreplyminetest.net works in gmail
18:19	celeron55	good
18:19	celeron55	Jordach, PilzAdam: you now can edit
18:20	Jordach	sweet.
18:21	celeron55	the template contains all the navigation menus and other things, so stuff about it can be added as issues on the github repo
18:21	BlockMen	hmmmm, could you post screenshot?
18:21	PilzAdam	oh god the page is slow
18:21	PilzAdam	celeron55, get a better internet connection
18:22	Jordach	PilzAdam, not slow for me
18:24	celeron55	and a policy on adding pages on the site: don't add pages that don't have direct navigation links at the top without first asking others - the goal is to keep that page as a front portal to minetest, and keep larger content in wikis and on the forum
18:25		proller joined #minetest-dev
18:26	Jordach	i just updated the current feature list
18:26	celeron55	i need to get a faster host for the site, but it's not something i can do in a short notice :P
18:26		Jordach joined #minetest-dev
18:26		Jordach joined #minetest-dev
18:27	celeron55	Jordach: the front page is missing all it's links, you might want to collect them back to it from minetest.net
18:27	Jordach	<Jordach> I changed the features bit to this: "Walk around, dig and build in an infinite voxel world (or boxel, as reddit calls it), and craft stuff from raw materials such as diamonds, MESE, Gold and other materials to build your dream world."
18:27	Jordach	celeron55, ^^
18:28	sapier	https://github.com/minetest/minetest/pull/636 Line of sight fix
18:28	Jordach	that's fast
18:28	sapier	was a silly error
18:28	sapier	los was just waste result of pathfinding ;-)
18:28	hmmmm	yeah, pilzadam, http://ompldr.org/vaTFqdQ
18:30	celeron55	that goes to BlockMen
18:30	PilzAdam	hmm?
18:30	sapier	lol ... commited wrong file only
18:31	hmmmm	oh
18:31	hmmmm	blockmen
18:31	hmmmm	i must've thought pilz asked me because their names have similar shapes and length, i didn't really read it, etc.
18:37	Jordach	anyways, you'll notice a huge difference from the old homepage to what i've modified
18:37	BlockMen	hmmmm, thx
18:38	sapier	can someone tell me how to drop something in latest minetest? :-)
18:38	Jordach	sapier, q key
18:39	sapier	thx
18:39	Jordach	sapier, you could have asked #minetest
18:39	Jordach	or even read the manual
18:39	sapier	lol yes I could but would have taken looooot of more time ;-)
18:39	Exio	this is funny
18:39	sapier	pilzadam how small is your rat?
18:39	Jordach	anyways, removing the nerd jargon helps new people understand the game better
18:40	PilzAdam	sapier, collisionbox = {-0.2, 0, -0.2, 0.2, 0.2, 0.2},
18:41	BlockMen	hmmmm, it is caused by the font. do you have zoom? or does someone else have that large font size?
18:41	thexyz	IceCraft: please, no away nicks
18:41	BlockMen	*zoomed
18:41	IceCraft	Yeah sorry
18:41	IceCraft	I will part until i fixed the problem
18:41	IceCraft	:/
18:41		IceCraft left #minetest-dev
18:41	VanessaE	bbl
18:42	PilzAdam	is that ok like this: http://c55.me/minetest4/doku.php?id=contributors
18:42	sapier	strange I just tried an animalmaterials egg wich is -0.12,-0.125,-0.12,0.12,0.125,0.12 it doesn't happen there
18:43	hmmmm	ahhhh
18:43	celeron55	PilzAdam: looks good
18:43	hmmmm	blockmen, i didn't really think that my font was really that big
18:43	hmmmm	but i see what's happening
18:44	celeron55	PilzAdam: altough... i don't really want my e-mail so exposed
18:44	hmmmm	there's nothing you can really do about it
18:44	Jordach	anyways, ive done some new stuff on the main page if anyone's interested
18:44	celeron55	i already get enough mails with the cryptic e-mail representation that currently is on minetest.net
18:45	celeron55	so change it somehow; maybe even remove altogether
18:45	PilzAdam	anyone else doesnt want the e-mail there: http://c55.me/minetest4/doku.php?id=contributors
18:45	sapier	there has to be another thing to happen too to cause this bug pilzadam
18:45	PilzAdam	^ hmmmm, thexyz, darkrose, RealBadAngel, proller
18:46	PilzAdam	sapier, they also have acceleration = {0,-10,0} and physical = true
18:48	PilzAdam	BlockMen, not all links in the header turn red when pointing with the mouse at them
18:48	sapier	no difference to egg pilzadam ... except beeing -9.81 but I doubt this makes a difference
18:49	thexyz	PilzAdam: i'd like some basic anti-bot protection
18:50	thexyz	like making captcha-style image displaying my email, then cutting it into pieces 1 pixel wide and inserting them into page
18:50	PilzAdam	ummm
18:50	sapier	but plz don't add one of those captchas driving even human ppl mad
18:50	PilzAdam	you can do that if you want
18:50	thexyz	well, it doesn't have to be hard to read
18:50	PilzAdam	Ill remove your e-mail for now
18:51	thexyz	i hope bots are too stupid to see page history
18:51	PilzAdam	isnt your e-mail on github too?
18:51	Jordach	thexyz, usually bots arent that smart, they look for keywords like email password new user, register
18:51	BlockMen	PilzAdam, i noticed that at Firefox only. and i have no idea why it does
18:52	BlockMen	someone want change? i just took the first ten mods with most hits http://c55.me/minetest4/doku.php?id=mods
18:52	thexyz	PilzAdam: it is, but it's not displayed anywhere
18:52	thexyz	apart from git history which bots don't download
18:52	thexyz	PilzAdam: well, I'm ok with keeping it
18:53	PilzAdam	it is in guiMainMenu.cpp
18:54	PilzAdam	thexyz, so I can add it back?
18:54	thexyz	sure
18:54	Jordach	anyways, im going to add nice new shiny screen shots with shaders on
18:54	thexyz	celeron55: are we going to keep that wiki-website at your server?
18:54	Jordach	(the old 0.4-dev look horrible)
18:55	PilzAdam	Jordach, screenshot in the main page is wrong
18:55	Jordach	fixing
18:56	Jordach	done
18:56	PilzAdam	you can remove IRC channels from there
18:56	sapier	hmmmm you should add the sky fix in order to fix the invalid emerge thread fix ;-)
18:56	PilzAdam	and authors too
18:59	thexyz	BlockMen: don't you think it should say "Most Popular" instead of "Recommended"?
19:00	proller	https://github.com/proller/common/commit/2dbfdeb4d10ec9964ed2a0713d30b949bc2cdc38 sky ores, springs
19:00	thexyz	I also don't get the capitalization there
19:01	thexyz	Description? Forum? Mods?
19:01		IceCraft joined #minetest-dev
19:01	BlockMen	thexyz, sry but my english isnt best...
19:01	IceCraft	Kay back
19:01	PilzAdam	sapier, *** glibc detected *** /home/adam/Minetest/minetest/bin/minetest: double free or corruption (!prev):
19:01	PilzAdam	happens every time I shutdown a server
19:02	thexyz	BlockMen: that's fine, mine isn't either
19:02	BlockMen	and it should be "Recommended" and not just a list of "Most Popular". I just dont know which to choose, so i took top 10 ;)
19:02	sfan5	will code invoked by minetest.after run in a seperate thread?
19:03	sapier	thats exactly what I meant with sky fix ;-)
19:03	thexyz	sfan5: all Lua is single-threaded
19:03	* sfan5	had hope for threading
19:04	sapier	sfan5 I'm working on it ;-)
19:04	BlockMen	PilzAdam, is there a reason you removed #mintest-doc at irc?
19:04	sapier	still it's a long way to go
19:04	PilzAdam	BlockMen, oops, forgot to add it
19:05	BlockMen	i edit already ;)
19:05	sapier	pilzadam do rats spawn automaticaly?
19:05	ShadowNinja	is the wiki at wiki.minetest.com still used?
19:06	PilzAdam	sapier, yes, but use /spawnentity mobs:rat
19:07	thexyz	ShadowNinja: apparently
19:08	thexyz	though that separation isn't good
19:08	ShadowNinja	Can I/how can I edit the navbar?
19:10	BlockMen	PilzAdam, could u add a comment? i dont know what this channel is used for...
19:10	ShadowNinja	/topic
19:10	proller	any objection about ? https://github.com/proller/common/commit/2dbfdeb4d10ec9964ed2a0713d30b949bc2cdc38
19:10	proller	when commit i make better message
19:11	ShadowNinja	abs height?
19:12	PilzAdam	sand into stone?
19:12	PilzAdam	and the liquids are way too rare
19:12	PilzAdam	I think it should be done in the c++ mapgen and not with register_ore()
19:13	proller	sky liquids ok, undergroud maybe rare, but i want to make smooth scarcity change
19:14	ShadowNinja	hmmmm has a commit for that, that he won't push...
19:14	celeron55	hmm, i think an "Overview" page is needed for "Community"
19:14	celeron55	i'll add that
19:14	proller	sand for sky lands and high mountains, there is no way to get it
19:14	celeron55	(to the navigation)
19:15	proller	and all sky changes doesnt affect standatd mapgen
19:16	BlockMen	celeron55, https://github.com/celeron55/minetest.net_dokutemplate/pulls
19:16	sapier	LOL Pilzadam your rats drop once you reach them in sky
19:19	proller	and make underground springs by config, or no? (i think no - nobody will disable it)
19:20	celeron55	ehm... what the hell happened
19:20		rarkenin joined #minetest-dev
19:21		rarkenin joined #minetest-dev
19:21	celeron55	oh, main.php got wrong permissions by git 8)
19:22	BlockMen	celeron55, fixed: &purge=true
19:24		emptty joined #minetest-dev
19:26	celeron55	BlockMen: oh by the way, are you sure you haven't accidentally left any page with PHP allowed?
19:27	Jordach	XSS ftl
19:28	celeron55	more like rm -rf
19:28	Jordach	heh
19:28	celeron55	quickly grepping, it looks like it isn't enabled anywhere
19:29	BlockMen	celeron55, im 100% sure.
19:30	BlockMen	and html is just allowed (by me) for support
19:31	PilzAdam	BlockMen, #minetest-doc isnt used by anyone anymore, Ill remove it
19:31		jojoa1997|Tablet joined #minetest-dev
19:31	ssieb	emptty: I don't have a repo set up, but you can get my current spec file from from http://sieb.net/minetest.spec
19:32	BlockMen	PilzAdam, ok
19:33	emptty	ssieb: cool, thanks. My debian package is here: https://github.com/mquinson/minetest-debian
19:33	emptty	ssieb: can I see your patches, too, pliz?
19:33	ssieb	oh, right
19:33	proller	PilzAdam,
19:34	proller	maybe generate springs only for indev mapgen ?
19:34	celeron55	Jordach: are you sure you didn't go a bit over the top with that "passion" stuff? 8D
19:34	PilzAdam	I add rules to IRC in the homepage, what should be added?
19:34	Jordach	celeron55, make it easy for the user to read = more users who speak english
19:34	Jordach	or other languages
19:34	ssieb	emptty: here's the src.rpm then: http://sieb.net/minetest-0.4.6-0.fc18.src.rpm
19:34	celeron55	Jordach: but i mean, the front page now has that word twice
19:35	Jordach	uh-oh
19:35	celeron55	also, why aren't there any trees or buildings in the front page screenshots? I'd imagine you could fit both of those in if you are going to have three of them
19:36	emptty	ssieb: did you include all games in the same package, or did you do separate packages for each kind of game?
19:36	Jordach	celeron55, well, im annoyed that i cant put the screenshots inline
19:36	Jordach	unlike pure html
19:36	Jordach	celeron55, IF the trees actually worked
19:37	celeron55	i think you can
19:37	Jordach	hold on, visiting redcrabs server for buildings
19:37	BlockMen	how bout that pic for front page? https://dl.dropbox.com/s/f6yh3brsjdxkww3/bone_02.png
19:38	celeron55	Jordach: just remove the : from the image tag
19:38	celeron55	umm...
19:38	celeron55	actually no; remove the |
19:38	PilzAdam	http://c55.me/minetest4/doku.php?id=irc
19:38	PilzAdam	anything missing?
19:38	celeron55	also, the styling seems to screw up with images floating on left side and a list on the right side
19:39	ssieb	emptty: I have minetest_game and common in it
19:39	celeron55	BlockMen: http://c55.me/random/2013-04/tscrot-2013-04-09_22-38-55.png
19:39	BlockMen	PilzAdam, "don't" instead of dont
19:39	* ssieb	is not the official packager, this is just for my use :-)
19:39	emptty	ubuntu does it too, I don't like this approach
19:40	emptty	what is the opinion of you guys, upstream? Will common and minetest_game evolve separately from the main source package ?
19:40	ssieb	well, it's new that there are multiple games, so it's definitely something to discuss
19:40	emptty	should we do separate packages in the distros, or not ?
19:41	BlockMen	celeron55, you mean that there is not enough margin to right?
19:41	Jordach	celeron55, got a nice picture from redcrabs server
19:41	emptty	I personally am very reluctant to modify the source package, but that may be only me. I'd like to have the advice of the Big Mens of this project to decide
19:41	celeron55	BlockMen: well, i'm not sure about what it's about, but there should be something that would cause some space between the image and the list
19:43	BlockMen	ok, i will fix
19:43	ssieb	emptty: what do you mean by modify the source package?
19:43	celeron55	emptty: the latest official windows package serves as the reference of the content of the "official minetest distribution"
19:44	celeron55	of course you're free to do differently, but that is what users expect
19:44	emptty	ssieb: include several source tarballs in only one source package
19:44	Jordach	celeron55, hows the current homepage now
19:45	celeron55	Jordach: if you're going to include an image where there is night, put at least a single torch down because it's useless to have an almost completely black image
19:46	PilzAdam	http://c55.me/minetest4/doku.php?id=texturepacks heh
19:46	ssieb	emptty: I don't see any problem with that.  But you could split up the games into their own packages, so e.g. the survival game could be its own package
19:46	PilzAdam	I would change "recommended" to "popular"
19:46	emptty	celeron55: why did you guys separate common and the games from the main tree? I'm puzzled about the mention in README that the minetest_game git might get reset
19:46	ssieb	emptty: I mean binary packages, the source package would have everything to generate them all
19:47	celeron55	ssieb, emptty: if you are going to split the package, make sure to make a metapackage called "minetest" that will install all of the packages, and call the bare engine + common something like minetest-engine
19:47	celeron55	emptty: common and the games are separated from the engine tree to keep the repositories reasonably sized
19:48	celeron55	common and games contain textures, sounds and 3D models which aren't wanted in the engine repo
19:48	emptty	is there any possibility that they get versioned differently?
19:48	emptty	like a 0.4.6.1 release of minetest_game while the other remain at 0.4.6 or so
19:48	celeron55	it is intended that they are identically versioned via git tags
19:49	Jordach	celeron55, anyways, forest screenie done
19:49	celeron55	it's complicated to keep up with it though now that there are so many of them...
19:49	Jordach	celeron55, hows that
19:50	celeron55	i think the city shot should be from daytime
19:50	Jordach	okay
19:50	celeron55	the dark trees shot is good
19:50	Jordach	(it's actually a town)
19:50	Jordach	the day desert one looks really real
19:50	BlockMen	celeron55, https://github.com/celeron55/minetest.net_dokutemplate/pull/2
19:51	emptty	celeron55: you mean that there will never be a separate release of any of the elements? I'm sorry but I need to know before I go for one or the other way of packaging these elements
19:51	PilzAdam	is there a global changelog for the dokuwiki like this one http://dev.minetest.net/Special:RecentChanges ?
19:53	Jordach	ShadowNinja, new image incoming
19:53	celeron55	BlockMen: hmm... i don't think that's good; i made a fix myself now that works well
19:53	BlockMen	PilzAdam, &do=recent
19:54	BlockMen	celeron55, whats wrong about my fix??
19:55	celeron55	it unnecessarily affected right-aligned images too, and the 1em margin was too low to actually move the list anywhere
19:56	* ShadowNinja	watches the main page change over and over
19:57	Jordach	celeron55, fixed the village image
19:57	Jordach	(rhyming today)
19:57	BlockMen	celeron55, and where do you see the change of the right-aligned images?? https://github.com/celeron55/minetest.net_dokutemplate/pull/2/files#L0L53
19:57	celeron55	hmm, there should be a link to ?do=recent somewhere... maybe in the account menu?
19:58	BlockMen	should i add?
19:59	celeron55	i think yes; it could be called just "Changelog" in there i guess
19:59	BlockMen	ok. and a general question first: how do i update my fork?
20:00	celeron55	there's probably a good answer to that somewhere on github itself :P
20:01	celeron55	this probably answers your question: https://help.github.com/articles/fork-a-repo
20:01	BlockMen	already found :D
20:02	Jordach	celeron55, for the recommended texture packs, should I spend time taking pics with each TP and make photos?
20:03		proller joined #minetest-dev
20:04	celeron55	i really think the first image on the front page should be a daytime screenshot with some houses in a forest
20:05	jojoa1997|Tablet	yeah that show the survival side of minetest
20:05	celeron55	Jordach: a screenshot for each sounds excessive; maybe split them into categories and take a screenshot for each?
20:05	celeron55	each category, i mean
20:05	celeron55	oh well, there aren't that many anyway it seems
20:05	jojoa1997|Tablet	caleron55 but each texture pack looks different
20:05	proller	celeron55, whats about placing server list (maybe top10 by players to bottom of main page ?)
20:05	Jordach	celeron55, the texture pack creators have made their own images, why bother doing more work
20:06	celeron55	Jordach: well just use them i guess :P
20:07	Jordach	uploading VanessaE's image
20:08	celeron55	just crop the images to similar size
20:08	celeron55	exact same width would be wise at least
20:08	celeron55	and not too much height
20:08	celeron55	oh lol it's awful now
20:08	celeron55	kill it with fire!
20:09	Jordach	celeron55, it can be done later
20:09	Jordach	content first, refinery later
20:12	Jordach	celeron55, fixed
20:15	emptty	"later" came quickly, this time :)
20:15	BlockMen	celeron55, my shell just throws me errors
20:15	BlockMen	what happens if i delete my repro and fork again?
20:15	PilzAdam	BlockMen, thats the way RealBadAngel does it all the time
20:16	BlockMen	PilzAdam, ok. then i will follow RealBaAngel :D
20:16	Jordach	RealBadAngel, is bad at branching out
20:18	Jordach	las image
20:18	Jordach	last*
20:18	ShadowNinja	given infinite resources how many players could connect to a minetest server?
20:19	Jordach	ShadowNinja, infinite provided that the server have enough ram
20:19	Jordach	although there was a time where a Intel ATOM notebook hosted 100+ people
20:19	PilzAdam	ShadowNinja, max_users in minetest.conf
20:20	celeron55	well, i guess some id there is 16-bit, so 65535 or so
20:20	celeron55	something's inevitably going to explode before that though
20:21	ShadowNinja	thats what I was asking, u16?
20:21	celeron55	i can't imagine anyone having the amount of single-core CPU power needed for that though
20:21	Jordach	celeron55, datacentres
20:21	celeron55	they have thousands of rather lazy cores
20:22	celeron55	not faster than one in a regular desktop
20:22	BlockMen	celeron55, https://github.com/celeron55/minetest.net_dokutemplate/pull/3
20:23	ShadowNinja	celeron55: why don't you just give him push access :-)
20:23	celeron55	why do you include an ID in there, it works like this too http://c55.me/minetest4/doku.php?do=recent
20:24	celeron55	ShadowNinja: because once i pull, all code will go to production immediately
20:24	BlockMen	because umm...
20:24	BlockMen	idk :D
20:24		jin_xi joined #minetest-dev
20:25	BlockMen	celeron55, changed
20:27	celeron55	now you could try learning a skill called "squashing"
20:27	celeron55	nobody will merge a pull request that has separate commits for fixing stuff that it introduces as broken 8)
20:27	ShadowNinja	celeron55: should a link to dev.minetest.net be added?
20:27	ShadowNinja	git rebase -i
20:27	Jordach	ShadowNinja, that's wise
20:28	PilzAdam	BlockMen, https://github.com/minetest/minetest/pull/605#issuecomment-15998442
20:28	celeron55	hmm, quite long https://help.github.com/articles/interactive-rebase
20:29	celeron55	PilzAdam: be careful
20:30	celeron55	there are three commits squashed in the example and nobody is telling BlockMen he needs to squash two
20:30	celeron55	everything will explode when he squashes three 8)
20:30	BlockMen	PilzAdam, celeron55 thx
20:31	BlockMen	celeron55, http://i.imgur.com/Q9l0UBD.png
20:31	BlockMen	it works for me with 2em
20:36	Jordach	celeron55, i added a basic community page
20:36	ShadowNinja	BlockMen: that was fixed, refresh
20:37	proller	https://github.com/proller/minetest/commit/02b6a1db212284412c036b2c674f6e4c3de59b7d - for fun - mapgen for math-generated worlds - first is simple sphere, next i want to make mandelbox
20:39	ShadowNinja	celeron55: all packs under Texture Packs does not start with a capital A
20:40	Jordach	ShadowNinja, im not seeing that
20:41	BlockMen	ShadowNinja, noticed. strange thing is, i had to refresh 10 times, and i was always forcing
20:41	Exio	proller: how does it work?
20:41	proller	Exio, makes simple sphere with r=30000
20:42	Exio	oh, it is not something like definition a function what returns the values in some math-like language? :P
20:44	proller	c++ langauge
20:45	Exio	hehe
20:47		iqualfragile joined #minetest-dev
20:48	ShadowNinja	Jordach: http://ompldr.org/iaTFsNQ
20:48	ShadowNinja	or better: http://ompldr.org/vaTFsNQ/Menu_019.png
20:49	ShadowNinja	ugh, why so many links?
21:00		jojoa1997|Tablet joined #minetest-dev
21:00		sapier1 joined #minetest-dev
21:05	sapier	https://github.com/minetest/minetest/pull/637 fix for pilzadams flyaway rats
21:06	sapier	I wonder why this only results in errors with small collisionboxes
21:25		emptty joined #minetest-dev
21:39		khonkhortisan joined #minetest-dev
21:39		emptty joined #minetest-dev
21:43		rarkenin|Away joined #minetest-dev
21:50		serengeor joined #minetest-dev
21:56	Taoki	PilzAdam: When can we expect sheep and rats to be included in Minetest with your simple mob framework? Can't wait to see that as part of minetest_game and Minetest builtin :)
21:59	sapier	I'm against this ... simplemobs is by far not best mob implementation
21:59	sapier	mine isn't perfect either but far more advanced than simplemobs
22:02	Taoki	I'm ok with any as long as there will be default mobs :)
22:02	PilzAdam	sapier, but simple mobs is lightweight enough to actually run on most peoples computers
22:02	Taoki	Simple might be more safe for Minetest at least for starters though
22:03	ShadowNinja	sapier: does mobf use pathfinding now?
22:03	Taoki	Pathfinding? *perks ears*
22:03	ShadowNinja	yep, it got pushed
22:03	sapier	simplemobs is by no means lightweight if you add half of features mobf has you'll require way more cpu power
22:03	Taoki	Was actually thinking of giving a try to pathfinding myself if no one else does. Since everything's made out of blocks t would be possible
22:03	hmmmm	hmm
22:03	hmmmm	sapier, what's the actual bottleneck with mobs
22:04	sapier	there's no actual benefit of pathfinding for about 95% of movement
22:04	Taoki	hmmmm - The only user who types his own nickname when he's thinking :)
22:04	sapier	current bottleneck is still spawning
22:04	sapier	and agression
22:04	hmmmm	how is spawning a bottleneck?
22:04	Taoki	sapier: Why would spawning be CPU costly? I was thinking movement would be hardest
22:05	sapier	no movement is simple
22:05	PilzAdam	sapier, yes, its lightweight because it doesnt have so many features
22:05	sapier	yes so your rat ignores water running submerged same as on ground ...
22:05	sapier	mobf 0.xx did same
22:05	Taoki	Only features I imagine are needed are mobs roaming around idly, then attacking the player (if hostile ones) when he's within their view radius
22:05		celeron55 joined #minetest-dev
22:06	sapier	spawning requires to check presence of other mobs as well as terrain thats costly
22:06	sapier	and agression is costly as you need to check what other objects are around
22:06	sapier	if you do this often you generate lots of cpu load if you do it to rare an agressive mob will attack with huge delay
22:06	sapier	both nit quite conveniant
22:07	PilzAdam	Taoki, simple mobs does exactly this
22:07	hmmmm	terrain checking?
22:07	Taoki	I recently wrote a moonflower mod (it's a plant) which spawns via mapgen with on_genrate. I used a very simply formula there
22:07	Taoki	PilzAdam: Awesome
22:08	hmmmm	first of all, you have control over all the other mobs, you're telling me that iterating through them and checking the distance from other mobs is cpu consuming?
22:08	sapier	mobf does only generate on mapgen since 2.0 ... simplemobs has 1/4th of mobs mobf has this is a big difference
22:08	hmmmm	maybe the spawn attempt rate should be lower
22:09	hmmmm	as for terrain checking... how do you perform that?
22:09	Taoki	Anyway, my formula (which I recommend is): For each chunk, have a given amount of probability attempts and possibilities for each. When a probability is triggered (via a math.random() happening), choose a random point within the chunk o the X and Z axes. Then at that spot, find the top on the Y axis and spawn there
22:09	sapier	the on mapgen method did reduce porblems to almost zero
22:09	PilzAdam	spawning is the most difficult part of mobs
22:09	sapier	still agression is a unsolved issue
22:09	PilzAdam	simple mobs abms are not the best solution
22:09	Taoki	So the steps are: For each cunk, trigger random probabilities. When they take place, find random X and Z location in that chunk. Then find the top by going up on Y at those coords. Then spawn there if appropriate
22:09	PilzAdam	and Id like to make the mobs not static, so they get deleted when the block unloads
22:09	sapier	taoki I've already done this
22:09	Taoki	If you can't spawn there the attempt is failed
22:09	Taoki	ok
22:10	Taoki	sapier: Surely shouldn't be CPU intensive in the slightest. Even a hand watch should be able to run such a code easily :P
22:11	sapier	if you happen to catch one of minetests performance eaters ... e.g. getnode you'll have no chance at all
22:11	Taoki	PilzAdam: Is spawning still difficult a formula like my idea?
22:11	sapier	an I've already told, mapgen spawning fixed the spawning problem ... but introduced a new one
22:11	hmmmm	is there not a get_ground_level() function already?
22:11	PilzAdam	Taoki, simple mobs dont spawn on generate
22:11	Taoki	sapier: With such an implementation, you'd only use getnode a few times every few seconds
22:11	PilzAdam	hmmmm, no
22:11	hmmmm	oh...
22:12	hmmmm	well maybe there should be
22:12	Taoki	PilzAdam: Oh, that's correct... thay shouldn't. Still the same idea can be used ona per-chunk basis
22:12	Taoki	Only horrible thing to do would be using register_abm on dirt nodes for this
22:12	sapier	taoki you don't have chunks if anythin is loaded
22:12	* Taoki	nods, makes sense
22:12	sapier	chunks are only a valid element on loading
22:12	PilzAdam	Taoki, but I have to register an abm on dirt nodes
22:13	sapier	as sheep won't spawn on stone or sand ;-P
22:13	PilzAdam	the dirt monster spawns on dirt
22:13	hmmmm	now tell me about the aggression part
22:13	Taoki	PilzAdam: It hurts to think how many functionc alls must be executed every X seconds for millions of dirt nodes loaded in view at once :P
22:13	sapier	ok atm agression needs to know any entity in a certain distance
22:14	sapier	and loop through this list in certain intervals ... e.g. every 1 5 10 seconds ... much more doesn't make sense
22:14	Taoki	PilzAdam: I'd suggest one of two things instead. 1 - Use a random position in the entire active area (part of the world that's loaded by the server), in case that's possible. 2 - If not, use player position and spawn mobs in a given radius from players. Shouldn't make a difference since if the player is away who cares if mobs spawn
22:14	Taoki	sapier: I suggested an idea for aggression once; Alert levels
22:15	Taoki	Not sure if I still have the thread...
22:15	sapier	alert levels do increas problem even more
22:15	Taoki	They would be a bit hard to implement, but prolly not performance intensive
22:15	sapier	as you need to update alert levels even for non agressive mobs this will result in permanent calculation for each mob
22:16	PilzAdam	Taoki, ehm... do you know the current API?
22:16	PilzAdam	(of Minetest)
22:16	Taoki	PilzAdam: Learning it during the last days
22:16	sapier	taoki loops are always performance critical
22:16	Taoki	Mostly yes for onw. Less for the lua entities part
22:16	PilzAdam	Taoki, what you says lets me think that you have no idea about it
22:16	Taoki	PilzAdam: How come? Where am I wrong?
22:16	sapier	"days"? either you are a genius or just haven't understood complexity ;-)
22:17	PilzAdam	<Taoki> PilzAdam: I'd suggest one of two things instead. 1 - Use a random position in the entire active area (part of the world that's loaded by the server), in case that's possible. 2 - If not, use player position and spawn mobs in a given radius from players. Shouldn't make a difference since if the player is away who cares if mobs spawn
22:17	PilzAdam	an ABM is exactly the latter one
22:17	Taoki	Oh, yeah. I doubt you can get a location globally in loaded chunks. Player position is possible to get in a global on_step however
22:17	Taoki	ok
22:18	sapier	hmmm I don't really see an option how to get attention mechanism faster ... maybe this isn't a problem with luajit anymore, but as I told, new performance evaluations will be done prior mobf 2.1 release (scheduled for end of april)
22:18	Taoki	PilzAdam: Still, how can the server even survive when you run ABM on EACH dirb block? :P I mean you can have almost million(s) of them loaded at once on the server in some circumstances I can imagine. At least thousands
22:19	sapier	it survives because pilzadam uses very very very low chances and huge deltas
22:19	PilzAdam	I check for neighbor = "air" and have a really low chance to run it
22:20	sapier	wich is the only way to go ... did you solve the everything cought/killed problem pilzadam?
22:20	PilzAdam	what problem?
22:20	sapier	with low chances and rates worlds tend to become empty if players kill mobs
22:21	PilzAdam	dunno, I use the engines damage/kill system
22:22	sapier	me to ... but thats not what I meant ... spawning at low rates doesn't seam to refill map at a decent rate
22:22	PilzAdam	bts, sapier, do you suppor the only_peaceful_mobs setting that is mentioned in minetest.conf.example?
22:22	Taoki	Oh. Checking for neighbors air is a great idea
22:22	Taoki	Should have used that for my Moonflower
22:22	PilzAdam	Taoki, it makes only sense in ABMs
22:22	Taoki	sapier: Huge deltas tho? What do you mean?
22:22	sapier	no wouldn't make sense as I do have at least 3 classes of mobs
22:23	sapier	completely peacefull, hostile and self defending
22:23	PilzAdam	Taoki, if you are interested, run simple mobs with display_mob_spawn = true in minetest.conf
22:23	sapier	but mobf has a configuration menu where you can disable each mob
22:23	Taoki	What would be wrong with using player position in on_step however? Any player's position, frequency multiplied by the number of players or something
22:24	Taoki	PilzAdam: Ok. Need to try the mod out soon yes
22:24	sapier	or try old mobf version on oldcoders server
22:24	PilzAdam	also, the engine passes the object count in the mapblock to the ABM function, I limit the mob spawn by this
22:25	sapier	object count in abm? interesting ... didn't know this
22:25	Taoki	Also, here's another idea which would also be nice gameplay-wise: Trigger mobs in register_abm very rarely, but have mobs which are spawned create mobs of the same type around themselves within a certain radius. So a mob of one type will cause more mobs of the same type
22:25	Taoki	Could be much less intensive also
22:25	PilzAdam	this isnt good at all, but it works in singleplayer
22:26	PilzAdam	Taoki, hmm, maybe that could work
22:26	PilzAdam	but if you clear an area, then the mobs will only very slowly come back
22:26	Taoki	Yeah. Could be minked to my idea of mobs that cry for help when fighting (and possibly low on health) causing more mobs to spawn nearby
22:26	PilzAdam	but I want something like the mob spawn of Minecraft
22:26	sapier	my current mechanism adds invisible spawners on mapgen wich produce mobs
22:26	Taoki	true too
22:27	Taoki	**linked
22:27	Taoki	sapier: Problematic if you build in that area and cover or uncover a spawner in the wrong place
22:27	PilzAdam	sapier, I thought about that too, but Im too lazy to code it
22:28	sapier	lol no it's not a problem taoki as spawner can detect this and relocate to a better position
22:28	Taoki	Hmmm. Might work then
22:28	Taoki	Yes... actually I do have an idea. When dirt nodes are created, give them a meta: field for mob spawning
22:28	sapier	hehe pilzadam ... thats what I meant with mobf is far more advanced ... I had to put a lot of work in those small glitches ... they seam to be minor but solution is quite tricky
22:29	sapier	a node doesn't have an active component ... otherwise it wouldn't be a node
22:30	Taoki	Anyway I might go inactive on IRC for now
22:30	Taoki	Yeah, true
22:30	sapier	pilzadam I like the mouse model of simplemobs ;-)
22:30	PilzAdam	its a rat
22:31	hmmmm	how does the mob spawning of minecraft work?
22:31	sapier	noo its a mouse :)
22:31	hmmmm	and how would you even know that
22:31	PilzAdam	hmmmm, they maybe use on_load_chunk() and check vertical position and time
22:31	hmmmm	so in other words no idea :)
22:31	PilzAdam	but I know how it feels
22:32	hmmmm	anyway i'll try to help out by making an env:ground_level_at_point() function
22:32	PilzAdam	on_load_chunk() would be extremly helpful if I delete mobs when the chunk gets unloaded
22:32	hmmmm	perhaps
22:33	hmmmm	and you mean on_load_block()
22:33	sapier	why do you delete mobs on unload?
22:33	PilzAdam	well, in MT block, in MC chunk
22:33	PilzAdam	sapier, I will do it when the patch gets in
22:33	sapier	not if but WHY?
22:34	PilzAdam	I dont like static mobs
22:34	PilzAdam	the player shouldnt be able to farm them
22:34	hmmmm	well i think being precise with naming here is important, because map is generated in chunks, and 5*5*5 blocks load all at once
22:34	sapier	I see so completely different aproach to what mobf tries to do
22:35	PilzAdam	hmmmm, I mean the 16x16x16 blocks
22:35	hmmmm	just keep that in mind
22:35	hmmmm	we'd have to call your function 125 times every time a chunk is generated
22:35	sapier	whooo :-) sonds costly
22:35	hmmmm	okay
22:36	hmmmm	i'm going to be honest
22:36	hmmmm	this is not for lua
22:36	PilzAdam	hmmmm, we need to call it everytime a player walks near to it
22:36	hmmmm	*sigh*
22:36	hmmmm	so i guess we go through this again?  i add a register_mob() function?
22:36	BlockMen	gd n8t everyone
22:36	sapier	what would this register function help?
22:37	hmmmm	spawning
22:37	sapier	and how to decide where to spawn? if you call a lua callback to tell all the slow parts will still be done in lua
22:37	PilzAdam	we need 2 types, one for static mobs and one for not static mobs
22:37	hmmmm	i guess i would pick a method and stick with it
22:38	sapier	imho you're solving the wrong problem atm
22:38	PilzAdam	it is the biggest problem for simple mobs currently
22:38	sapier	your problem is getting a suitable position?
22:38	PilzAdam	the problem is to find a good spawning system
22:39	hmmmm	i thought the problem was making new mobs evvvvery time a new block is loaded
22:39	hmmmm	which i personally don't agree with, i think it should be done once after map generation, and then mobs are added every so often afterwards pseudorandomly
22:39	PilzAdam	hmmmm, thats a solution, not a problem
22:39	sapier	why do you wanna waste cpu performance on every block load instead of only once?
22:39	hmmmm	that's not a solution, just consider the sheer volume of callbacks
22:40	hmmmm	it's a solution to you because you don't like static mobs for some reason
22:41	PilzAdam	it would be also possible to add a "ghost" entity when a mob dies that regenerates nearby after some random time
22:41	sapier	I think about a parametrizable spawn mechanism in core but my investigations aren't quite finished ... it's difficult to get this done in a way general enough without having to call back to lua everytime
22:42	hmmmm	hah
22:42	hmmmm	what if mobs were placed with decorationdef
22:42	sapier	this won't fix the to many objects problem where core just deletes entities
22:43	sapier	I was talking about ghosts
22:43	PilzAdam	hmmmm, could work
22:43	PilzAdam	we would have static mobs then
22:43	PilzAdam	I dont like it but it isnt the worst thing one can do
22:43	sapier	what do you mean with decorationdef?
22:44	hmmmm	it's just a generic way to randomly place objects on the map after map generation
22:44	hmmmm	by objects, i originally figured it would be discrete node structures like trees and rocks
22:45	sapier	does it call a lua fct for random positions?
22:45	hmmmm	it comes up with the position on its own
22:45		khonkhortisan joined #minetest-dev
22:46	hmmmm	it has a lua callback if it wants one though
22:46	PilzAdam	sapier, your object collision fix works, hmmmm, can I merge it?
22:46	sapier	if you did try it yes ... it worked for me
22:46	hmmmm	basically i'm thinking like, you register it almost like you do an ore, but a couple other differences, and you specify which biomes have what decorations
22:46	hmmmm	pilzadam, yes
22:46	PilzAdam	yes, I tested it
22:47	hmmmm	and one of the fields i'd have is a lua callback
22:47		khonkhortisan joined #minetest-dev
22:47	hmmmm	so this way you can place cacti in deserts according to perlin noise but not have any lua callback
22:47	sapier	sounds good
22:47	hmmmm	but for something like mobs now
22:47	hmmmm	hmm
22:47	hmmmm	i guess place CONTENT_IGNORE, which means nothing, but do a callback to lua for that random position
22:48	sapier	at least a lua function has to be called for any "possible position" to decide if a specific mob really fits there
22:48	hmmmm	(this random position is really just random x and z, but ground level y, by the way)
22:48	PilzAdam	hmmmm, the engine writes to errorstream if you place content_ignore
22:48	hmmmm	well obviously
22:48	hmmmm	that could be a placeholder though for "don't place anything at this position, just call a lua callback for it"
22:49	sapier	onground isn't good enough for all mobs
22:49	hmmmm	well for those mobs, you could do something different i suppose
22:49	sapier	birds, fish, lavamonsters ;-)
22:49	PilzAdam	hmmmm, then just pass nil as param
22:49	sapier	later one haven't been done by now :-)
22:51	hmmmm	pilzadam, you misunderstand me.. i just blurted out a minor implementation detail
22:51	hmmmm	you'll see how it'll work when i get time to work on it
22:51	sapier	I don't like the idea of adding another solution only solving half of the problems ... if I need to do everything else with same way as before the problem isn't solved
22:51	PilzAdam	I know
22:51	PilzAdam	I just wanted to be sure that you dont do something stupid in the details
22:51	hmmmm	do i ever?
22:51	hmmmm	why would you even ask that
22:52	sapier	btw hmmmm are you currently working on mapgen v6?
22:52	hmmmm	nope, proller is working on that
22:53	sapier	after running valgrind on current master I feel mapgen got even more memory leaks than before
22:53	hmmmm	if you're referring to the noiseparams thing, i'll get all of those at once
22:53	PilzAdam	sapier, after all your memory leak fixes minetest crashes every shutdown with a double free error
22:53	sapier	speaking of memory leaks I rebased the sky patch to current master and added the fix for current minetest crash in there too
22:53	PilzAdam	and it also crashes if a mod fails to load
22:54	sapier	as I have said 3 times now this fix is included within sky patch
22:55	sapier	pilzadam same problem with mod load errors
22:55	hmmmm	sapier, i'll take care of all mapgen related memory leaks.. just not right now
22:55	hmmmm	it's not exactly critical either
22:55	hmmmm	you guys are all in such a hurry for some reason
22:55	sapier	no it isn't but should be done prior releasing 0.4.7 imho
22:56	hmmmm	we *just* released 0.4.6
22:56	hmmmm	that won't happen for a while
22:56	sapier	I just guess it won't be done at all if not someone puts his finger in this wound ;-)
22:57	sapier	but at least the big texture leaks are fixed now, once everything is correctly added this will be way better than before
22:57	proller
22:58	PilzAdam	sapier, will test your line_of_sight() fix tomorow, gtg now
22:58	PilzAdam	bye
22:58	sapier	me too it's 1 am
22:59	sapier	good night .. afternoon morning ... wherever you are :-)
22:59	hmmmm	well sapier
22:59	hmmmm	hold up
22:59	sapier	yes?
22:59	hmmmm	just want to make sure, all of the mapgen-related memory leaks are related to noiseparams, right?
22:59	VanessaE	meh, scrollback tl;dr.
22:59	hmmmm	there's nothing else that i'm not aware of
23:00	sapier	not quite but I'll post a report once all of my fixes are in to show whats leftover
23:00	sapier	I think I've seen one or two new ones today but I'm not sure I didn't especially look for those
23:02	sapier	but the sky fix is important, it's anoying to crash minetest on each logout :-( sorry for that
23:04	sapier	reason for including it in sky is its a one line fix I didn't want to create an additional pull request for 1 line ;-)
23:04	sapier	so gtg now good night
23:04		sapier left #minetest-dev
23:10	RealBadAngel	shit, found bug in my code
23:27	VanessaE	bbl
23:45		Taoki joined #minetest-dev

← Previous day | Channels | #minetest-dev index | Today | Next day → | Google Search | Plaintext

Powered by ilbot.