IRC log for #minetest-dev, 2012-12-17

Time	Nick	Message
00:53	hmmmm	well, the answer depends on the weird message.
01:13	VanessaE	wouldn't "weird message" be pretty much the standard for most programs' errors? :-)
04:08	thexyz	It's "access violation in trans_func" or something like this
04:20	thexyz	happens when i push_back into that list
06:36		jin_xi joined #minetest-dev
07:51		jnumm joined #minetest-dev
08:10		kotolegokot joined #minetest-dev
08:20	jnumm	Hi!
08:20	jnumm	I was updating the Launchpad PPA packaging
08:21	jnumm	and Lintyian gave warnings about a few things
08:21	jnumm	*Lintian
08:22	jnumm	* manpages use "-" but they should use "\-"
08:22	jnumm	* a spelling error: succesfully => successfully
08:23	jnumm	I have patches for them at http://bazaar.launchpad.net/~minetestdevs/minetest-c55/packaging/files/head:/patches/
08:23	jnumm	Would someone mind having a look?
09:08		SpeedProg joined #minetest-dev
10:12		Calinou joined #minetest-dev
10:17		rsiska joined #minetest-dev
13:27		rsiska joined #minetest-dev
13:52		rsiska joined #minetest-dev
13:57		SpeedProg joined #minetest-dev
14:14		doserj joined #minetest-dev
15:28		SpeedProg1 joined #minetest-dev
15:37		PilzAdam joined #minetest-dev
15:39		Calinou joined #minetest-dev
15:57	thexyz	so, i have core::list<MediaFetchThread> m_media_fetch_threads; and when i m_media_fetch_threads.push_back(MediaFetchThread(this)); access violation occurs; if i use std::list instead of core::list everything works fine
15:58	celeron55	why do you push a thread to a container by using the copy constructor?
15:58	celeron55	8)
15:59	celeron55	doesn't sound good
16:01	thexyz	how should i do it then?
16:01	RealBadAngel	hi all
16:02	celeron55	store pointers (or smart pointers) to MediaFetchThreads allocated on the heap (and remember to delete them in an appropriate destructor)
16:03	RealBadAngel	celeron55, can you say watch this? and say what are you thninkin bout it please: http://www.youtube.com/watch?v=GocjjOLuogk
16:03	celeron55	an extra good thing about that is that you can remove MediaFetchThread from the header and just declare the class name in there
16:03	RealBadAngel	ive made L-systems tree, by now in LUA
16:05	thexyz	celeron55: also, did you read our conversation about stl/irrlicht containers?
16:06	celeron55	19:53:53 < thexyz> hmmmm: why is irrlicht's core::list used everywhere then?
16:07	celeron55	because initially minetest was coded solely using irrlicht's things
16:07	celeron55	and i have a bad habit of never getting into mass-changing things like that
16:08	thexyz	ok, so such change would be accepted
16:09	celeron55	yes
16:10	RealBadAngel	celeron55, what about mine?
16:14	celeron55	there are some related things i need to note: when using stl's containers, iterators should be incremented using ++i, not i++ (for perforamnce reasons) (currently the latter is used because some of irrlicht's ones don't even implement the former); also, take into account that g++'s implementation of std::list has an O(N) performing .size(), so the appropriate cases (comparisons to 0) should be changed to .empty()
16:14	celeron55	i don't even know what irrlicht does, but anyway...
16:15	celeron55	also on the list of stupid random things: in game.cpp, the user input is checked using a huge bunch of else ifs instead of ifs for no reason
16:15	celeron55	it doesn't cause anything in practice but somebody probably wants to change it too 8D
16:15	RealBadAngel	celeron55. are you willin to respond or not?
16:16	celeron55	RealBadAngel: are you willing to wait or not?
16:16	thexyz	celeron55: thanks
16:16	RealBadAngel	im waitin
16:16	thexyz	i've read about std::list::size but already forgot it
16:17	celeron55	RealBadAngel: those trees are boring compared to the effort i suspect you have put into it
16:18	RealBadAngel	it took me a few days to dig for all needed libraries
16:19	RealBadAngel	remade matrix4 functions in lua
16:19	celeron55	why not make the trunk thinner and make it branch earlier?
16:19	RealBadAngel	program them
16:19	RealBadAngel	its a language
16:20	RealBadAngel	L-systems
16:20	RealBadAngel	and you call them boring?
16:20	celeron55	i don't care about how you've done them; i'm mainly interested in interesting results :P
16:20	RealBadAngel	well, fyi community call it awesome
16:21	Calinou	"interested in interesting results" -- how can you be interested in something not interesting?
16:21	darkrose	RealBadAngel: just give him a pony-shaped tree, might shut him up for a bit
16:21	celeron55	Calinou: dunno, people seem to require that sometimes
16:21	RealBadAngel	let him program his own tree
16:22	RealBadAngel	lsystems are able to make "celeron" tree with turtle commands
16:22	thexyz	celeron55: http://bpaste.net/show/64985/ is it ok now? have i messed something up again?)
16:22	RealBadAngel	its a universal plant generator
16:22	VanessaE	celeron55: to be fair, his trees are, as the word goes, epic.
16:23	Calinou	isn't it slow?
16:23	celeron55	it's no reason to not try to be better
16:23	VanessaE	the mod is buggy but it works as a proof-of-concept - I tried it locally to very good visual effect
16:23	PilzAdam	the trees are epic compared to the current ones, but there is much more potential in it
16:23	VanessaE	Calinou: quite so.
16:23	RealBadAngel	it needs matrix4 functions
16:23	RealBadAngel	lua sucks with it
16:23	VanessaE	it needs the meat of it programmed in C++
16:23	thexyz	RealBadAngel: probably, luajit can help in this case
16:24	RealBadAngel	ive translated even irrlicht native funcrtions to lua for this mod to worrk
16:24	Calinou	thexyz is the 61st person to ask for luajit support.
16:24	thexyz	and the first one who tried it with minetest and found it was actually slower, right?
16:24	Calinou	(we should have it, there, 62nd :P)
16:24	RealBadAngel	i dont wont luajit
16:24	Calinou	why would it be slower?
16:24	celeron55	thexyz: you should probably create the threads in the client's constructor
16:25	RealBadAngel	im almost done with c++ version
16:25	thexyz	Calinou: well, i trusted my profiler that time
16:25	thexyz	celeron55: sure
16:25	celeron55	thexyz: that way they could end up being created twice if stuff changes
16:25	thexyz	Calinou: it probably lied then
16:25	Calinou	thexyz: did you try it with eg. a mobs mod?
16:25	Calinou	or snow mod (lots of lua mapgen)?
16:25	celeron55	thexyz: otherwise it looks like the way things are already done in other places
16:26	RealBadAngel	celeron55, one question. basic. have you seen the video i posted?
16:26	thexyz	Calinou: i don't remember (probably not, try searching forum for my profiler output), build minetest with luajit and try it yourself
16:26	celeron55	RealBadAngel: ehm, i wouldn't have answered anything if i didn't look at it first
16:26	Calinou	thexyz: doesn't luajit support require a lot of additional C++ code, or is it just a compiler option to enable?
16:26	RealBadAngel	so you think its just "boring"???
16:28	thexyz	Calinou: the easiest (and hackiest) way to build minetest with luajit is to build minetest and luajit separately, then replace liblua.a with libluajit.a, then `make` minetest again, it'll link with brand new luajit instead of lua
16:28	RealBadAngel	its really, really sad to hear that again and again
16:28	RealBadAngel	good night
16:29		kotolegokot joined #minetest-dev
16:30	thexyz	celeron55: http://bpaste.net/show/64986/, fine now?
16:31	celeron55	i mean, there are many kinds of trees: http://2.bp.blogspot.com/_h0NrIKmjyhM/TH-L7M6cuAI/AAAAAAAAGoI/ScW1UBEto5U/s1600/P9020002.JPG http://www.vastavalo.fi/albums/userpics/16126/normal_Pihlaja.jpg http://www.helsinki.fi/metsatieteet/arboretum/images/pinus_sylvestris_puu.jpg http://bhubbard.smugmug.com/Other/Finland-After-Midsummer/i-6PzTwJL/0/L/img_0017-L.jpg ...
16:31	celeron55	... http://virtuoosi.pkky.fi/metsaverkko/metsan_mittaus/puun_tilavuus/kuvat/muotoluvuttehtava.gif http://s3.freefoto.com/images/15/19/15_19_8_prev.jpg
16:31	celeron55	you're just making the last kind of those random examples
16:32	celeron55	well, maybe the third too, a bit
16:32	thexyz	celeron55: btw, in "realtest" game there are multiple kind of trees
16:32	celeron55	but if you're going to make them so big, you should try to really seek the limits of what can be made with nodes
16:33	jin_xi	i quite like the scale
16:34	celeron55	RBA seems to have some kind of a problem towards improving good things
16:35	thexyz	http://i.imgur.com/rSw0D.png
16:35	PilzAdam	thexyz, realtest uses predefined shapes IIRC
16:36	thexyz	PilzAdam: they're randomized a bit
16:36	thexyz	better ask kotolegokot though
16:37	kotolegokot	what is IIRC
16:37	PilzAdam	if i remember corretly
16:37	celeron55	(you're supposed to google those without nobody knowing you didn't know 8))
16:38	kotolegokot	ok
16:38	thexyz	or use irc logs, those are deciphered in there
16:42	celeron55	also, i think i'd like big forests rather than individual huge wide trees
16:44	jin_xi	ok, so i hope some rambling is ok if you don't mind. But i want it up in the logs that I applaud RBAs efforts. Generating trees with some structure is not easy and his results are very interesting.
16:47	celeron55	sure
16:52	VanessaE	agreed.
16:55	thexyz	celeron55: ok, i've tested and new one works fine; will push it shortly
16:55	thexyz	thanks for your assistance and everything
16:59		rubenwardy joined #minetest-dev
17:05	PilzAdam	I guess I can push this: https://github.com/PilzAdam/minetest/commit/f4ed37ec6f1d24c2da3dab14b3478d3b0452d7aa
17:06	VanessaE	and I have to say this, also for the logs: celeron55, stop calling everything boring et.al. What's boring to you may be interesting to the rest of us, and vice versa. You want contributors? STOP PUSHING THEM AWAY.
17:06	VanessaE	don't read too much into that, take it at face-value only.
17:07	celeron55	can i call you boring? 89
17:07	celeron55	8)'
17:07	celeron55	*
17:07	VanessaE	:P
17:08	thexyz	they take it too seriously
17:09	celeron55	PilzAdam: i am not sure if it's the right place but if it works, it's fine enough
17:17		rubenwardy left #minetest-dev
17:21		hmmmm joined #minetest-dev
17:22		rubenwardy joined #minetest-dev
17:28		loggingbot joined #minetest-dev
17:28		Topic for #minetest-dev is now Minetest core development and maintenance. Chit-chat goes to #minetest. Consider this instead of /msg celeron55. http://irc.minetest.ru/minetest-dev/ http://minetest.net/wiki/doku.php?id=dev_index
17:38	VanessaE	anyway... celeron55, RBA's trees are interesting. they're certainly a far cry from the trees the game has now. Will you implement his L-systems engine or not?
17:38	VanessaE	Forget the tree axioms he chose, just the engine is important right now
17:38	doserj	Does someone want to comment on https://github.com/doserj/minetest/commits/nodeboxtype_connectable, before I continue with it? It's a quick start to try to make nodes with dynamic nodeboxes like fences, pipes, etc easier to define.
17:39	doserj	it's untested and incomplete, but I would like to know if I am on the right way
17:41	PilzAdam	whats about texturing?
17:41	PilzAdam	different textures for the connection boxes?
17:42	doserj	I don't think that#s easy to do
17:42	celeron55	VanessaE: i won't, but as long as it works for a larger variety of tree types, i'm fine with including it
17:42	VanessaE	that makes no sense?
17:42	celeron55	"Will you implement"
17:42	VanessaE	"I won't, but... i'm fine with including it"??
17:42	VanessaE	ok you mean you'll add his code, but not write it yourself.
17:43	celeron55	that's not called implementing
17:43	VanessaE	sorry, wrong choice of word.
17:43	VanessaE	and actually in American slang, to "implement" can also be to use someone else's tools to do something,.
17:44	VanessaE	but anyway
17:44	hmmmm	erm, "approve"
17:44	celeron55	in programming, it simply means coding something
17:44	hmmmm	he approves for it to be implemented
17:44	* VanessaE	shrugs
17:45	VanessaE	let's not argue over semantics.
17:45	celeron55	you're the one arguing here about you yourself using wrong words 8)
17:45	hmmmm	it seems like the l-system code is quite compact
17:46	VanessaE	grumble
17:46	hmmmm	should be easy
18:04	thexyz	celeron55: http://irc.minetest.ru/minetest-delta/2012-12-17#i_2728628
18:07	celeron55	umm
18:09	celeron55	i'm too tired to try to undestand what is going on
18:09	celeron55	but it doesn't sound good
18:09	PilzAdam	would you mind when we add a new parameter to register_on_placenode()?
18:10	celeron55	what parameter
18:10	PilzAdam	itemstack
18:10	thexyz	celeron55: you mean, it can be out of sync?
18:10	celeron55	thexyz: i think we must preserve the possibility of eg. dual-wield or whatever
18:11	PilzAdam	the main problem is that modders cant modify the itemstack in on_placenode()
18:11	thexyz	ok
18:11	PilzAdam	so just pass it to the function and everything will be fine
18:11	thexyz	PilzAdam: feel free to implement this then
18:13	celeron55	looking at minetest.item_place_node, there are some nontrivialities
18:14	celeron55	i guess you are able to solve them 8)
18:22	PilzAdam	both methods work: https://gist.github.com/4320583
18:23	thexyz	PilzAdam: where's diff of builtin?
18:23	PilzAdam	wait a sec.
18:24	PilzAdam	https://gist.github.com/4320598
18:24	* PilzAdam	wonders what nontrivialities c55 was talking about
18:25	thexyz	well, that one is wrong
18:26	thexyz	it will be tricky to add, say, 500 items when stack_size is 99
18:26	thexyz	you should modify itemstack, then modify player's inventory, then hope that everything will go right
18:31		rsiska joined #minetest-dev
18:37	PilzAdam	add a return value wether items should be taken after the callback call?
18:37	thexyz	i prefer things to "just work"
18:37	thexyz	so inventory modification should "just work"
18:38	thexyz	(that's what my patch achieves, except i don't know is it safe to rely on wield item or not)
18:38	thexyz	many mods already do, afaik
18:43	PilzAdam	https://gist.github.com/4320598
18:45	PilzAdam	what do you think?
18:46	celeron55	the first callback call is a bit error-prone, just make it like the second
18:50	PilzAdam	everyone is fine with this? https://github.com/PilzAdam/minetest/commit/29c17c14f1ef42f8080a44b1e3308f34de8b5df2
18:52	celeron55	it looks robust, but i am not aware of everything something might do 8)
18:52	VanessaE	mmmm...
18:52	VanessaE	I don't know if returning "true" is necessarily right though
18:52	VanessaE	shouldn't it return false, as if to say "this didn't work"?
18:52	PilzAdam	its actually "returning not false"
18:52	celeron55	of course not, it should default to automatically taking the item
18:52	celeron55	it's all about defaults
18:53	celeron55	even if it wasn't for compatibility
18:53	VanessaE	shrug
18:55	PilzAdam	Ill just push it and we will see
18:55	celeron55	"break early, break often" or something along those lines 8-)
18:56	VanessaE	hah!
18:56	VanessaE	then push my mese changes too as long as we're breaking mods :D
18:58	PilzAdam	so lets modify creative now: infinite stacks of size 1 and not drop digged items in the inventory
18:58	celeron55	is that the way people like it to work?
18:59	PilzAdam	its actually the MC way
18:59	celeron55	i think it might be useful to get the item that one dug, if one doesn't already have it in the inventory
18:59	celeron55	to easily re-place it
18:59	PilzAdam	thats a good idea
19:00	PilzAdam	also modify the hand so it breaks everything in 0.1 seconds?
19:00	VanessaE	nooooo
19:01	VanessaE	that will cause too much stuff to be dug with each action,
19:01	PilzAdam	ok
19:01	VanessaE	try it with a maptools admin pick sometime
19:01	celeron55	does it have the problem of continuing the digging if the mouse is held down?
19:01	celeron55	if it does, then it's not good
19:01	PilzAdam	maybe 1 second
19:02	PilzAdam	so you dont have to use tools
19:02	VanessaE	0.25 seconds.
19:02	PilzAdam	nah, 0.5
19:02	PilzAdam	deal?
19:02	VanessaE	that's just a bit longer than the amount of time a mouse click would last
19:02	celeron55	i think people like the feel of digging things
19:03	PilzAdam	creative is all about building fast
19:03	VanessaE	let's compromise: 0.375 seconds :-)
19:04	celeron55	i think it should be made so that in creative mode, a mouse click will remove a single node immediately, and while the button is held down, no more things are dug until it is clicked again
19:04	VanessaE	celeron55: but only with the hand?
19:04	PilzAdam	i tried that but thats really weird
19:04	celeron55	you get used to it
19:05	VanessaE	celeron55: if a tool can cut faster than that, the tool should still cut faster.
19:05	PilzAdam	removing over 5 blocks is just annyoing with it
19:05	VanessaE	have you ever stood at the side of a hill with an admin pick? you can raze the whole thing in a couple of minutes if you have good coordination, just holding the button down.
19:06	celeron55	oh well, whatever; let's say if somebody plays minetest creative multiple hours a week for some other reason than testing, that one can choose
19:06	PilzAdam	VanessaE, mese pick has 0.5 seconds for stone and thats fast enough
19:06	VanessaE	no, it isn't.
19:06	celeron55	...
19:06	celeron55	make it 0.4
19:07	PilzAdam	you still need to be able to open doors
19:07	celeron55	THE END.
19:07	celeron55	8D
19:07	VanessaE	PilzAdam: roll mouse wheel, click door, roll wheel back.
19:07	PilzAdam	if the hand is that fast?
19:07	VanessaE	I'm talking about the speed of tools now, not the hand.
19:08	VanessaE	do whatever you want with the hand if it works better,
19:08	VanessaE	but don't disturb how tools work.
19:08	PilzAdam	sure
19:08	VanessaE	in that case, 0.4-0.5 seconds recoil time I guess is okay for the hand
19:11	PilzAdam	crumbly, cracky, snappy, choppy, oddly_breakable_by_hand; do I miss any group?
19:27	PilzAdam	https://github.com/PilzAdam/minetest_game/commit/61e60724f6b0875d14dc7f185d914702eddb806e
19:27	PilzAdam	celeron55, VanessaE, this is basically the implementation of what we talked about
19:29	VanessaE	ok, I can see that being about right then
19:30		doserj left #minetest-dev
19:31		doserj joined #minetest-dev
19:36	PilzAdam	so what do you think? push it or not?
19:40	hmmmm	if we add enchanting like in minecraft you can make efficiency tools you can instamine
19:41	hmmmm	which would be very useful for terraforming
19:41	hmmmm	also we need a diamond shovel/axe/sword
19:41	celeron55	oh god, here is the word
19:41	hmmmm	errr, mese
19:42	celeron55	do you have any idea what terraforming means? 8)
19:42	hmmmm	forming large pieces of natural land into the shape that you'd like
19:43	hmmmm	i used it appropriately http://en.wikipedia.org/wiki/Terraforming
19:44	celeron55	i don't see how modifying planets to be earth-like is related to anything you might do
19:44	celeron55	i can't recall the actual correct word for the thing you mean though
19:44	VanessaE	landscaping
19:45	celeron55	that sounds good
19:45	celeron55	it's completely hilarious how the minecraft community has started using the word "terraform" for landscaping
19:45	celeron55	it's just... terribly wrong
19:45	VanessaE	isn't the whole MC community just "terribly wrong"? :-)
19:46	celeron55	it almost makes me cry :-D
19:47	PilzAdam	can anyone comment on this: https://github.com/PilzAdam/minetest_game/commit/61e60724f6b0875d14dc7f185d914702eddb806e ?
19:49	sfan5	https://github.com/celeron55/minetest/pull/359
19:51	sfan5	can anyone comment on that?
19:52	celeron55	thexyz probably wants to
19:53	PilzAdam	I guess due to we talked about this creative thing I can just merge it
19:53	thexyz	sfan5: indentation is messed up
19:54	sfan5	where?
19:54	thexyz	everywhere!
19:54	thexyz	https://github.com/sfan5/minetest/blob/ce5f73760ef001c5cfd33e9174d4589222c18dbb/cmake/Modules/FindCURL.cmake
19:54	sfan5	oh...
19:54	thexyz	and i'm not really good at cmake, what's the difference, anyway?
19:55	thexyz	you didn't need to close it
19:58	sfan5	thexyz: is this good? https://github.com/sfan5/minetest/blob/f40171f738d29c1745c91c73768127a541a64b3a/cmake/Modules/FindCURL.cmake
19:58	thexyz	it's consistent, that's enough for me
19:58	thexyz	so, care to explain changes?
20:00	sfan5	that makes searching for curl libs/headers easier on unix and it also prints the determined variables
20:00	sfan5	https://github.com/celeron55/minetest/pull/360
20:03	sfan5	so.. will this get merged?
20:03	celeron55	i wonder why it doesn't work without it
20:03		jin_xi joined #minetest-dev
20:04	PilzAdam	https://github.com/PilzAdam/minetest_game/commit/95cbfc50d8f7fc492f6def60720e355fdd0f62c1
20:04	celeron55	it's not like cmake wouldn't search in /usr/local/include and /usr/include otherwise
20:04	celeron55	and the others in there
20:05	celeron55	is it now in line with other custom Find*.cmake files that we have?
20:05	celeron55	or was it previously?
20:05	celeron55	i guess it should be similar to them
20:10	PilzAdam	https://github.com/PilzAdam/minetest_game/commit/95cbfc50d8f7fc492f6def60720e355fdd0f62c1 <- comments?
20:12	thexyz	PilzAdam: wait, you broke this by your previous commit?
20:12	thexyz	which changed stairs: to modname:
20:12	PilzAdam	yep
20:13	PilzAdam	because its calls modname in the after_place_node functions
20:13	PilzAdam	and that cant get the modname correctly
20:14	PilzAdam	now it works correctly
20:14	thexyz	then merge it
20:16	thexyz	let's hope it haven't enough time to break many mods
20:16	thexyz	hasn't*
20:16	thexyz	damn, i'd better go sleep
20:16	thexyz	celeron55: please, comment something on sfan5's pull https://github.com/celeron55/minetest/pull/360
20:32		rubenwardy left #minetest-dev
20:41	sfan5	thexyz: celeron55: cross-compiling fails when using curl, it doesn't link curl correctly: http://pastie.org/private/odron3rm1cahp0qei3b45q
20:41	thexyz	sfan5: it works fine for me
20:42	thexyz	you're probably trying to link static curl to minetest
20:42	thexyz	that won't work
20:42	thexyz	use dynamic version
20:42	thexyz	what was the point of asking the question and leaving?
20:51	thexyz	almost forgot it, celeron55, anything else about my .ttf patch?
20:51	thexyz	i think it's safe to merge because it's disabled by default
22:13	* hmmmm	shoots himself
22:14	hmmmm	I can't seriously believe how stupid i am
22:14	hmmmm	the latest mapgen bug that i couldn't find for 2 days was because I was accessing an array in a transposed manner
22:15	hmmmm	for (x ... ) { for (z ...) { instead of for (z ...) { for (x ...) {
22:15	celeron55	finding the reason makes you quite not stupid 8)
22:16	celeron55	the stupid ones are the ones who can't fix their own bugs
22:17	hmmmm	wow, I am rather pleased with my mapgen's output... now I just need to add in the interpolation between biomes to account for the height differences...
22:17	hmmmm	regardless it's still not going to be the default mapgen until maybe 0.5.0 if it's up to me
22:18	hmmmm	right, so my current task is to class-ify the old one and clean it a bit
22:28	hmmmm	haha... christ.... EmergeThread: mapgen lighting update: 0.6623 EmergeThread: mapgen::make_block: 0.6762
22:29	hmmmm	I think i know what needs fixing next
22:29	hmmmm	when 95% of the makeChunk time is being consumed by lighting updates...
22:33	hmmmm	actually 98%
22:35	VanessaE	ouch.
22:45		rsiska joined #minetest-dev

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