Minetest logo

IRC log for #minetest-dev, 2014-10-02

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

All times shown according to UTC.

Time Nick Message
00:24 DuDraig joined #minetest-dev
00:52 Miner_48er joined #minetest-dev
01:15 Zeno` joined #minetest-dev
02:05 NakedFury joined #minetest-dev
02:22 prozacgod joined #minetest-dev
02:23 mos_basik joined #minetest-dev
03:05 CraigyDavi` joined #minetest-dev
03:19 rickmcfarley joined #minetest-dev
03:45 Hunterz joined #minetest-dev
04:50 kaeza joined #minetest-dev
04:51 T4im joined #minetest-dev
05:33 hmmmm ugh
05:33 hmmmm you know what, I'm going to unify name resolution
05:33 hmmmm the way it happens right now works but it's sloppy
05:34 hmmmm have your data structure (biomes, decorations, schematics, whatever have you) include a std::string or a list of strings of the node name, then on emerge manager init, resolve them to the ID for each and every instance of these resources
05:36 hmmmm i'm going to expand INodeDef to keep a list of strings to content_t *s, or std::vector<content_t> *s (if the node string name represents a group) and when the node registration finishes, that will carry out the assignments
05:37 hmmmm this way, everything is unified, saves on the potentially huge space of extra std::strings that don't need to be maintained after name resolution, etc.
05:42 rickmcfarley joined #minetest-dev
05:43 VanessaE um guys?  why are there still old "minetestc55" builds on the "daily-builds" and "stable" PPAs?
05:43 VanessaE that build is horribly outdated
05:45 Hunterz joined #minetest-dev
05:48 hmmmm because obvs older software == more stable
05:52 VanessaE older all the way back to Ubuntu Natty though?  eh...
05:55 OldCoder joined #minetest-dev
05:57 Zeno` I'm having a problem: https://github.com/minetest/minetest/pull/1707
05:58 Zeno` I cannot *nicely* check that the destination slot is either contains the same item or is empty and not move(drop) if that's not true
05:59 Zeno` mainly because I'd have to duplicate the first 6 lines in the next conditional or make those lines a function or do something else kind of messy (I think)
06:02 Zeno` I suppose I could add another bool
06:18 hmmmm wouldn't the modulo logic you already have be better expressed as a bool like "m_drop_item"
06:18 hmmmm and then you just do m_drop_item = !m_drop_item instead of m_lmouse_move_mode = (m_lmouse_move_mode + 1) % 2;
06:19 hmmmm and you say "first 6 lines in the next conditional" - where is that you're talking about?
06:20 Zeno` Yeah I originally kept track of button releases as well, so maybe a bool with ! as you suggested is fine
06:21 Zeno` in the main code (full file)
06:21 Zeno` lines...
06:21 Zeno` 3146-3151
06:22 Zeno` I think
06:23 kahrl or even, instead of a bool, a set of slots that the item has already been dropped into
06:23 kahrl oh nvm
06:23 hmmmm hmm
06:23 kahrl thought you could initialize that with the initial slot but that would be ugly
06:23 hmmmm Zeno`, code it up the 'messy' way and we'll see what we could do from there
06:23 Zeno` kahrl, I was trying to mimic the MC behaviour where you can right drag around to evenly distribute the item but NOT swap with an item in a slot that contains a different item
06:24 Zeno` ok I'll do it the messy way
06:24 hmmmm I personally like the MC behavior
06:24 hmmmm so +1 for that
06:24 hmmmm you have my le upboat
06:24 kahrl Zeno`: oh, in that case add the set I mentioned (maybe in addition to the bool/int/whatever)
06:24 kahrl so every slot can only be dropped into once
06:24 kahrl (that's how it works in MC)
06:26 Zeno` I need to find the video someone showed me in that case :)   searching logs
06:27 kahrl I really think at this point the inventory handling in GUIFormSpecMenu has become way too overgrown
06:27 kahrl it has been like that for a long while already
06:27 kahrl it should be split into a separate class
06:28 kahrl (but don't bother about it in this fix)
06:28 hmmmm guys, what's the consensus on legacy hard-coded ABMs
06:28 hmmmm never going to happen again?
06:28 hmmmm what is the performance impact of growing grass from within Lua?
06:28 VanessaE hard-coded in the engine?  FUCK no
06:29 hmmmm some things like growing grass might be better off in the engine though.
06:29 VanessaE I have enough of an issue with ABMs in the default game; growing grass from Lua isn't that big of an impact.
06:29 hmmmm that's not an issue
06:30 hmmmm you could add a tiny knob in lua to turn that on or off
06:30 VanessaE after all, you only hit that kind of code once every, what, 5 or 10 minutes?
06:30 hmmmm that code gets hit very often
06:30 VanessaE well wait, you mean the dirt <-> grass ABM?
06:30 hmmmm yes
06:30 VanessaE eh
06:30 VanessaE that's not a big deal
06:30 hmmmm but how do you knooow that
06:31 VanessaE because Nostalgia game manages it without even a hiccup
06:31 hmmmm erm.. whatever
06:31 kahrl the probability is high enough that I don't see a problem
06:31 kahrl err *low
06:31 VanessaE I had to backport that change when that ABM was moved out of the engine and I saw no performance impact at all
06:32 Zeno` kahrl, I was just talking about that to someone else heh
06:33 hmmmm we want to stay envlocked for as short a period time as possible
06:33 hmmmm going back and forth from native to lua thousands of times over and over is a hit
06:33 hmmmm one that's definitely not necessary
06:34 hmmmm you say it's not much of a hit, but have you tried to quantify it at all?
06:34 hmmmm if we can save 10ms, that's awesome
06:34 Zeno` messy way is pushed (I hate it)
06:34 Zeno` I doubt it behaves exactly like MC, but it makes sense (well, to me anyway)
06:35 VanessaE hmmmm: at the time, of course, there was no profiler, so I only had my CPU stats page to go by ( http://digitalaudioconcepts.com/va​nessa/hobbies/minetest/stats.html ).  Nostalgia was a very busy server then, and it barely registered on the CPU graph.
06:35 kahrl hmmmm: would it be possible to only do the C++ -> Lua transition once for all registered ABMs? (per abmhandler.apply)
06:36 hmmmm kahrl:  wait, is that how it works??
06:36 kahrl hmmmm: not at the moment I don't think
06:36 hmmmm yeah I figured as much
06:36 hmmmm ugh
06:36 kahrl hmmmm: but it could be done
06:36 hmmmm UGHH
06:36 hmmmm let's do it
06:36 hmmmm as if the server step isn't slow enough already
06:39 kahrl it was done this way so that the C++ code can ensure a later called ABM still has satisfied conditions after the earlier ones ran
06:39 kahrl wonder how much would break if that wasn't done
06:41 hmmmm 'satisfied conditions'?
06:42 kahrl i.e. node is in nodenames, certain neighbors exist, etc.
06:43 kahrl oh and it recalculates the active object count if an ABM added any
06:44 hmmmm yeah I saw that much
06:45 hmmmm well look, ABMHandler's ctor calculates all the chances and trigger contents there
06:46 hmmmm so aside from the active object count, it can't possibly do what you said it did before
06:46 hmmmm the only thing it does in between ABMs is check the active object count
06:47 hmmmm can that piece of logic be moved outside of each ABM call for that block??
06:48 kahrl oh you're right, it doesn't handle every case
06:48 kahrl it rechecks the content after moving on to the next node though
06:48 kahrl (the outer loop in apply())
06:49 kahrl but if a node changes and there are more ABMs for that node, it only rechecks the neighbors
06:49 kahrl so I guess ABMs already have to deal with the chance the node might actually not be the node they were assuming it was
06:50 kahrl how many ABMs actually use active_object_count/_wider?
06:51 hmmmm not many... i can't think of any practical utility for that
06:51 kahrl perhaps it could be removed entirely
06:51 VanessaE kahrl: in my ABMs I've never actually done that unless I was checking for more than one kind of node
06:51 VanessaE (chance of a node not being the one it's assumed to be)
06:52 VanessaE I've never used the *object* features myself, but I'll bet sapier does.
06:52 kahrl VanessaE: in most cases the player wouldn't notice anyway (as the two ABMs happen in the same server step, so it wouldn't matter)
06:52 hmmmm hmmmmmmmmmmmm.
06:52 kahrl VanessaE: but it could break something if node metadata is involved
06:52 hmmmm kahrl, I'm thinking of a situation where two different mods would use something like a sapling
06:52 hmmmm one would grow it into a certain kind of tree and one into another
06:52 kahrl hmmmm: yeah, that too
06:52 hmmmm well the result would be a hybrid
06:54 hmmmm this is tough
06:55 kahrl VanessaE: I think mobf uses way more complicated algorithms than checking active_object_count
06:56 VanessaE kahrl: nope
06:56 VanessaE he uses it
06:56 hmmmm so I don't get it
06:56 kahrl hrm
06:57 VanessaE hm, seems I use it in moretrees' leaf decay too
06:57 kahrl simplemobs does some stuff with it too
06:57 VanessaE (likely as result of some copypasta from default)
06:57 hmmmm if mods already assume their ABM nodes are really the ones they wanted, and we don't see breakage already, then what would the harm be?
06:58 VanessaE vanessa@rainbird:~/Minetest-related/mods$ grep -ir active_object_count *|wc
06:58 VanessaE 174    1216   20940
06:58 VanessaE I'm sure most of those are false positives.
06:58 kahrl VargaD: add a grep -v function in there
06:58 VanessaE heh
06:58 kahrl oops, sorry
06:58 kahrl VanessaE: *
06:59 VargaD hi
06:59 kahrl hi ;)
07:00 VanessaE hmmmm: I can't say that there'd be any; don't most mods assume their ABMs will execute more or less independently?
07:00 hmmmm they do yeah
07:01 hmmmm no mod has a guarantee that another mod didn't change their node that had an ABM triggered on it.  no mod has a guarantee that another ABM didn't change a neighboring node
07:01 hmmmm well, the latter assertion isn't true right now, but i'd like that to be true so we could just do one lua call for every abm
07:01 VanessaE that's what I figured.  so then to fix it (so to speak) so that that assumption is in fact enforced shouldn't break anything at all
07:02 VanessaE if some modder plays off the idea that their ABMs will magically occur at just the right moment, they're insane
07:03 hmmmm indeed
07:03 hmmmm these kinds of glitches that the way ABMs work right now to prevent are practically unavoidable
07:03 hmmmm hrmm it's 3 AM
07:03 hmmmm lol I thought I was going to get a commit in!
07:03 hmmmm guess not
07:03 kahrl heh
07:04 VanessaE ok, ...
07:04 VanessaE kahrl: he DOES use active_object_count.
07:04 kahrl VanessaE: who's he?
07:04 VanessaE sapier I mean.
07:04 VanessaE just in one place though, in tracing.lua
07:04 kahrl hmm ok
07:05 hmmmm guys, what do you think about moving the declaration for CNodeDefManager out into nodedef.h
07:05 kahrl hmmmm: why?
07:06 hmmmm reduce a level of indentation
07:06 hmmmm cleaner
07:06 kahrl it does expose implementation details though and slows compilation
07:07 hmmmm well
07:07 Zeno` it only seems to be used in nodedef.cpp
07:07 hmmmm I'll keep it in
07:07 VanessaE kahrl: 12 pseudo-legit refs in my archives, after filtering out "action = function".
07:07 hmmmm I get the whole opaque implementation of an abstract interface idea but I don't think it matters that much in practicality
07:08 hmmmm it'd be a huge improvement to just get rid of that extra indentation by separating the method definitions and the class declaration, though.
07:08 kahrl hmmmm: you could also split class declaration and function implementation but keep them both in nodedef.cpp
07:08 Zeno` Well I pass by pointer or reference 99% of the time so I forward decls work fine. I think it's pretty much a muchness though
07:08 hmmmm yeah that's just what i said.
07:08 kahrl hmmmm: oh, thought you still meant moving it to nodedef.h
07:10 kahrl 86a3c8ce56 was a recent commit that did something like that
07:19 hmmmm am I the only one who doesn't like crap like this
07:19 hmmmm if (f->name == "default:water_source")
07:19 hmmmm is_water_surface = true;
07:20 VanessaE oh no..
07:20 VanessaE where are you finding that?
07:20 hmmmm that is a target for api-ization once client side lua becomes a thing
07:20 hmmmm VanessaE, CNodeDefManager::updateTextures()
07:21 hmmmm it looks like the work of RBA
07:21 hmmmm I don't blame him though, it's the least bad option available currently
07:21 VanessaE I wouldn't touch that if I were you, he'll kick your ass ;)
07:23 VanessaE hm
07:24 VanessaE yeah.  I haven't seen the code he's working on but if I guess it right, that glistening surface shader is gonna end up messing with this
07:24 VanessaE he wants to keep the water node name hard-coded :-/
07:25 VanessaE *looks at clock*
07:25 * VanessaE pokes RealBadAngel
07:25 VanessaE and no, I don't like it either.
07:25 VanessaE it should depend on some kind of flag/field in the node def
07:26 hmmmm like I said though, it's the least bad option
07:26 hmmmm given the restrictions and so on
07:26 hmmmm but yeah.
07:26 VanessaE I suppose so
07:26 hmmmm client side lua couldn't come fast enough.
07:26 hmmmm people would be able to distribute shaders with their mods
07:26 VanessaE I guess it could be worse:  it could be using a content ID instead.
07:26 hmmmm so their special node can ACTUALLY be special
07:27 VanessaE now THAT's a good idea
07:27 hmmmm ehe
07:27 hmmmm all this is a lot of talk though
07:27 VanessaE talking helps plan things out :)
07:27 hmmmm I'd be able to do half this crap if I take a vacation from work
07:27 kahrl I don't know if client side lua would actually mean mod-defined shaders...
07:27 hmmmm or maybe I just need to double down on coding
07:28 kahrl remember all the security issues webgl had?
07:28 hmmmm yeah
07:28 kahrl and still has probably
07:28 VanessaE kahrl: good point.  a shader API then?
07:28 kahrl VanessaE: maybe
07:28 hmmmm that was fixed up at the driver level though
07:28 VanessaE I mean, a sandbox, you know
07:29 hmmmm I don't think web browsers did anything to secure webgl
07:29 kahrl http://security.stackexchange.com/question​s/13799/is-webgl-a-security-concern#13840
07:29 hmmmm it's certainly a DoS concern
07:29 kahrl that post is a bit old though
07:30 hmmmm yeah but
07:30 hmmmm mods can't steal information this way because it wouldn't be able to recontact the server with the data
07:30 hmmmm that'd be very non-trivial, anyway
07:31 hmmmm also minetest servers have a slightly higher degree of trust than a random webserver
07:31 ImQ009 joined #minetest-dev
07:32 VanessaE hmmmm: that's debatable ;)  but your point stands
07:33 proller joined #minetest-dev
08:47 ImQ009 joined #minetest-dev
08:47 jin_xi joined #minetest-dev
09:03 jin_xi shader security concerns? im curious, because thats what i did (local hack): make shaders per nodedef passable/possible
09:04 BlockMen joined #minetest-dev
09:05 BlockMen going to merge https://github.com/minetest/minetest/pull/1694 in an hour
09:06 kahrl BlockMen: have you settled the disagreement with SN over whether to remove the preload items cb?
09:07 kahrl jin_xi: well, if it's a local hack then it's not much of a security concern :3
09:08 BlockMen kahrl, he said "Nah, that's O.K."
09:10 kahrl I see it now
09:10 jin_xi surely not. its just never ocurred to me that security is an issue with shaders
09:10 exio4 joined #minetest-dev
09:10 jin_xi or can be
09:11 kahrl BlockMen: is the case in nodedef supposed to fall-through?
09:11 kahrl BlockMen: (if it is, it definitely needs a comment)
09:12 kahrl I mean this: https://github.com/minetest/minetest/pull/1694/​files#diff-70868aa6d6b96c0c1623c761500d23c4R670
09:14 BlockMen oh, i missid adding content_mapbloc.cpp
09:14 BlockMen its handled like with allfaces_optional
09:15 Zeno` if that's the case that's not what it's doing
09:18 BlockMen kahrl, updated
09:19 edlothiol joined #minetest-dev
09:22 kahrl Zeno`: what do you mean?
09:23 kahrl BlockMen: connected_glass should be in minetest.conf.example
09:28 BlockMen kahrl, added
09:29 kahrl BlockMen: I think it's good to go then
09:29 kahrl merge it
09:30 Zeno` Is is being handled by NDT_ALLFACES or NDT_ALLFACES_OPTIONAL ?
09:30 kahrl BlockMen: wait, the case is back to fall-through
09:31 BlockMen Zeno`, no?
09:31 BlockMen what make you think that?
09:31 Zeno` because of the missing break; it looks like it's handled by NDT_ALLFACES
09:33 Zeno` (handled in part)
09:33 BlockMen wtf, indeed
09:34 kahrl huh, I thought for sure you fixed it in the earlier update
09:34 kahrl am I just imagining things?
09:35 BlockMen maybe it got lost with a rebase i did
09:35 BlockMen idk
09:36 BlockMen ok, everyone look again plz
09:36 BlockMen updated
09:37 kahrl BlockMen: looks good now
09:41 BlockMen im waiting for travis and then push it
09:42 rickmcfarley left #minetest-dev
09:58 kaeza joined #minetest-dev
10:03 Salbei left #minetest-dev
10:04 Zeno` going to merge https://github.com/minetest/minetest/pull/1707 in one hour
10:26 kaeza joined #minetest-dev
10:29 proller joined #minetest-dev
10:38 ninnghazad joined #minetest-dev
10:40 ninnghazad anybody else noticed that minetest.* functions that rely on getScriptApiBase do not work properly when called from a coroutine within lua? wondering if this might be intentional
10:40 ninnghazad getScriptApiBase returns a different stack (?) when called from a coroutine, and that seems to mess up the functions, as they read/write on the wrong indices then.
10:42 RealBadAngel BlockMen, for glasslike framed you will need different glass textures
10:42 RealBadAngel its frames are regular 3d boxes
10:42 RealBadAngel holes there looks terrible
10:43 BlockMen RealBadAngel, imo not but if more will complain it will be changed
10:43 RealBadAngel i suggest using wood and obsidian textures for frames and separate glass surface texture
10:44 RealBadAngel like so: https://github.com/minetest-technic/​framedglass/blob/master/init.lua#L39
10:45 dirkk0 joined #minetest-dev
10:45 dirkk0 left #minetest-dev
10:46 RealBadAngel BlockMen, theres already an issue posted bout it
10:46 RealBadAngel https://github.com/minetest/minetest/issues/1695
10:47 kaeza joined #minetest-dev
10:47 BlockMen RealBadAngel, he claims it should be rotated different
10:47 BlockMen not that there are gaps at all
10:49 RealBadAngel hes wrong
10:49 RealBadAngel i just closed that issue
10:49 RealBadAngel box is solid. needs separate texture, so you do too
10:50 proller joined #minetest-dev
10:54 RealBadAngel wise is to bound wood and obsidian textures to it, when user will modify textures for them corresponding glass will also change
10:55 RealBadAngel and btw, it cannot be rotated different way
10:56 RealBadAngel rotations are the same for all the boxes
10:56 RealBadAngel and no matter how you will rotate it there will always be a gap
10:57 RealBadAngel texture is made so
10:57 BlockMen ^ you should add that to the closed issue ;)
10:58 RealBadAngel adding irc log link
10:58 ImQ009 joined #minetest-dev
10:58 Amaz joined #minetest-dev
11:00 BlockMen k
11:09 RealBadAngel BlockMen, but imho you should change game usage of it
11:09 RealBadAngel the way i mentioned already
11:10 RealBadAngel in my mod you can find rdy textures
11:10 proller joined #minetest-dev
11:13 Zeno` Ok, I am now merging the commit
11:13 RealBadAngel otherwise some1 is gonna open similar issue again
11:13 Zeno` Ok, I think I found a problem
11:14 Zeno` I don't have permissions
11:14 RealBadAngel Zeno`, what commit?
11:14 Zeno` 1707
11:15 Zeno` and yes there is a yucky part in that... but hmmmm suggested I do it the yucky way
11:15 ninnghazad create #1709 for the problem with coroutines/minetest.*/stack
11:17 RealBadAngel Zeno`, how it behaves when both stacks sum is greater than max ammount?
11:17 Zeno` it should behave the same way as before
11:18 Zeno` I didn't change that logic
11:20 RealBadAngel if so im ok with it
11:26 RealBadAngel also im about to close all open issues/prs made by me
11:27 RealBadAngel all the code will be aviable only in my fork
11:27 RealBadAngel mt is free to take it
11:27 RealBadAngel but i wont be making any single pr again
11:28 RealBadAngel im done with that
11:28 kaeza joined #minetest-dev
11:29 kahrl 2015: every (ex-)minetest dev has their own fork
11:29 Amaz XD
11:30 RealBadAngel whatever. making your code into core takes more time than coding actually
11:30 RealBadAngel and is extremaly frustrating
11:30 kahrl eh, mine take months to be merged as well but I don't rage quit
11:30 PilzAdam joined #minetest-dev
11:30 RealBadAngel im not quitting
11:30 kahrl essentially you are
11:31 Zeno` I don't think it would be good losing your PRs
11:31 Zeno` sigh
11:31 RealBadAngel if you will decide to pick up my code i can help
11:31 RealBadAngel but i wont be begging for it anymore
11:32 ninnghazad understandable
11:32 kahrl well, I don't beg for mine getting merged either
11:33 RealBadAngel i just want my own code be up to date all the time
11:33 RealBadAngel without constanly need for rebasing
11:33 RealBadAngel which started to be just sick
11:34 kahrl if you intend to merge changes from minetest you will have to fix those conflicts instead
11:34 kahrl so it doesn't really help
11:34 RealBadAngel there are no conflicts here
11:34 kahrl so what's the problem with rebasing?
11:35 RealBadAngel i am coding new things based on previous changes
11:35 kahrl just say in the pull request, "depends on #xxxx"
11:36 RealBadAngel mapblock mesh in its current state is unusable for me
11:36 kahrl I've done that often actually and it worked out fine
11:37 RealBadAngel i tried to change it step by step
11:37 RealBadAngel but all my changes are blocked
11:38 kahrl blocked? more like no one's around I think
11:38 RealBadAngel for weeks?
11:38 kahrl yeah, I don't know why
11:38 RealBadAngel lately i had 2 weeks off
11:38 kahrl I suggested the weekly meeting but it never happened
11:39 RealBadAngel planned to spent that time for coding
11:39 RealBadAngel i havent managed to get a single pull merged
11:40 RealBadAngel yesterday i was just pissed off
11:40 RealBadAngel today im calm
11:40 RealBadAngel but im moving with my toys
11:41 kahrl I do agree that someone else should client/audiovisuals maintainer because c55 is barely around these days
11:41 kahrl should be*
11:42 RealBadAngel nobody is takin care of it
11:42 RealBadAngel in fact the project is slowly dying
11:42 kahrl open an issue about it on github and ping every dev to comment on it
11:43 RealBadAngel theres no progress at all
11:43 RealBadAngel and all the good things are shut down at pr stage
11:44 kahrl judging by the page hits recently posted I don't think it's dying
11:45 RealBadAngel you think prs changing spelling or mouse behaviour are progress?
11:45 kahrl why not?
11:45 RealBadAngel its stagnation
11:46 kahrl polishing a game is actually often more work than coding the first prototype
11:46 kahrl but necessary to make a good game (or engine)
11:46 RealBadAngel ofc but only when actively coding
11:46 RealBadAngel otherwise we end up like mojang did
11:47 kahrl I don't think microsoft would be interested in us
11:47 RealBadAngel with fireworks as new shiny feature
11:47 RealBadAngel or new kind of wood
11:47 RealBadAngel its called dead end elsewhere
11:48 proller changing fetchresult to fetch_result was very important feature.
11:49 kahrl if the engine is already powerful enough that good progress (in your sense) can be done in mods, I don't see the problem
11:50 RealBadAngel kahrl, its not
11:52 diemartin joined #minetest-dev
11:52 kahrl I see at least a couple dozen new mods release in the last few weeks
11:53 kahrl of course, if you see any individual mod as progress is up to you, but those are the numbers
11:53 RealBadAngel thats just an effect of mojangs actions
11:54 RealBadAngel for a short period of time we will be able to see folks trying to contribute
11:54 RealBadAngel but it will end as usual
11:54 kahrl and your fork will help this how?
11:55 RealBadAngel i just want to code, its my hobby
11:55 RealBadAngel and i will do that all the time
11:55 hintss joined #minetest-dev
11:55 RealBadAngel i just dont want anymore crawling and begging for it
11:55 RealBadAngel its pointless
11:56 kahrl again, just don't beg for it and it will be less frustrating
11:56 RealBadAngel its gone too far
11:57 kahrl RealBadAngel: part of the problem is that there are 91 open pulls requests. how are devs supposed to randomly pick yours to comment on?
11:57 kahrl pull*
11:58 RealBadAngel just grep the logs and see how many times i asked for comments on my pulls
11:58 jp_ joined #minetest-dev
12:00 ninnghazad still, having to look at large pulls is time-consuming at least. and the guys having to do those probably also just want to code, and not have to read other people's code.
12:01 ninnghazad i'd say don't take it personally
12:01 RealBadAngel so the development style just suck balls
12:01 kahrl RealBadAngel: how often do you merge pull requests of others?
12:02 kahrl I don't keep logs but I don't think I've often seen you do that
12:02 kahrl so look in the mirror please
12:02 RealBadAngel actually i do that quite often
12:02 RealBadAngel ask zefram_fysh
12:03 RealBadAngel hes a main contributor for technic last days
12:03 jp_ RBA : sent you a PM
12:04 RealBadAngel i just trust him and his code
12:04 RealBadAngel and hes doing great job
12:04 blaaaaargh joined #minetest-dev
12:04 kahrl RealBadAngel: well, I'm not talking about technic
12:04 RealBadAngel its the same
12:05 RealBadAngel technic mod is one of the biggest mods for mt out there
12:05 ninnghazad when he breaks something that you've missed to spot, and BILLIONS of angry miners come at you for it, you shall change your mind!
12:05 jp_ left #minetest-dev
12:06 RealBadAngel thats the point
12:06 ninnghazad all like "oh noes, why didn't i spend 27 hours checking those PRs..."
12:06 RealBadAngel code needs to be verified
12:06 RealBadAngel by players
12:06 ninnghazad hm, well... no... yes... also maybe but not primarily i think
12:06 RealBadAngel our current model of development seems to forget bout it
12:07 RealBadAngel whats important for devs are whitespaces
12:07 RealBadAngel or style issues
12:07 RealBadAngel but not the features
12:07 proller and hyper compare in player save
12:07 ninnghazad well that sucks. sounds like coding @work ;)
12:08 kahrl so me complaining about a missing "break;" is about style issues?
12:08 RealBadAngel lol
12:08 RealBadAngel you know what im talkin about
12:09 RealBadAngel and missing break is not a style issue but a bug
12:09 RealBadAngel lately i pointed missing break too
12:10 RealBadAngel in firelike drawtype
12:18 ninnghazad which PR are you waiting on btw?
12:20 blaaaaargh since we are talking about drawtypes, I wish modders could specify per-cuboid textures for nodeboxes
12:21 kahrl https://github.com/minetest/minetest/issues/1710 see, I even wrote this issue for you RBA
12:21 kahrl (any dev welcome to comment on that.)
12:21 kahrl anyone else of course too
12:22 Zeno` I'd support RBA
12:22 Amaz
12:23 RealBadAngel blaaaaargh, no need for that with mesh drawtype aviable
12:24 RealBadAngel that makes literally nodebox drawtype obsolete
12:26 indriApollo joined #minetest-dev
12:28 kaeza probably. nodeboxes are easier to "edit" though, if you don't have experience with 3D modelling software
12:28 Amaz
12:29 kahrl fun fact: we could simply stop supporting nodebox
12:29 kahrl "nodebox -- See below. EXPERIMENTAL" -- lua_api.txt, Oct 2014
12:29 kahrl "Please note that this is still experimental, and may be incompatibly changed in the future."
12:30 kaeza or simply s/EXPERIMENTAL//, etc
12:30 kaeza it has been "EXPERIMENTAL" for what, 2 years?
12:30 kahrl heh yeah I didn't think of that
12:30 kahrl yes, sounds right
12:33 ninnghazad2 joined #minetest-dev
12:35 ninnghazad_mobil joined #minetest-dev
12:37 ninnghazad_mobil joined #minetest-dev
12:38 ninnghazad_mobil joined #minetest-dev
12:40 proller joined #minetest-dev
12:40 ninnghazad_mobil joined #minetest-dev
12:45 ninnghazad_mobil joined #minetest-dev
12:47 ImQ009 joined #minetest-dev
12:59 shadowzone joined #minetest-dev
13:02 indriApollo joined #minetest-dev
13:44 Taoki[mobile] joined #minetest-dev
14:00 diemartin joined #minetest-dev
14:14 ninnghazad_ joined #minetest-dev
14:15 blaaaaargh joined #minetest-dev
14:19 zat joined #minetest-dev
14:38 ninnghazad1 joined #minetest-dev
14:45 hmmmm joined #minetest-dev
14:53 ninnghazad_ joined #minetest-dev
14:56 ninnghazad1 joined #minetest-dev
15:02 Megaf_ kahrl: Are you there?
15:04 Megaf_ kahrl: are do you mean by Build taker?
15:06 ninnghazad joined #minetest-dev
15:06 ninnghazad_ joined #minetest-dev
15:07 ninnghazad1 joined #minetest-dev
15:07 Megaf_ so, is there a plan to make oficial builds?
15:07 Amaz Official builds for what?
15:07 Megaf_ ^ sfan5
15:08 sfan5 Megaf_: for what?
15:08 Megaf_ https://github.com/minetest/mine​test/issues/1710#issue-44679910
15:08 Megaf_ Item 4
15:08 ninnghazad_ joined #minetest-dev
15:08 sfan5 Megaf_: "Build" is the building system, not minetest builds
15:08 sfan5 Megaf_: btw: I already do the MinGW release builds
15:08 sfan5 the msvc ones are done by BlockMen
15:09 Megaf_ oh, I see
15:09 Megaf_ I could make minetest builds...
15:10 sfan5 why would you?
15:10 sfan5 there are enough
15:10 sfan5 for windows at least
15:11 diemartin joined #minetest-dev
15:14 Megaf_ I'd like to make a generic Linux build
15:15 sfan5 then make a static build
15:15 sfan5 that however doesn't work because some glibc functions need that version of glibc at runtime
15:16 sfan5 "warning: Using 'getaddrinfo' in statically linked applications requires at runtime the shared libraries from the glibc version used for linking"
15:17 ninnghazad joined #minetest-dev
15:19 ninnghazad_ joined #minetest-dev
15:43 blaaaaargh joined #minetest-dev
15:58 PenguinDad joined #minetest-dev
15:59 ninnghazad joined #minetest-dev
16:01 mos_basik joined #minetest-dev
16:08 Calinou joined #minetest-dev
16:12 rubenwardy joined #minetest-dev
16:20 BlockMen joined #minetest-dev
16:21 blaaaaargh joined #minetest-dev
16:24 dulrich left #minetest-dev
16:26 MinetestForFun joined #minetest-dev
16:32 BlockMen sfan5, you need to update your builds. -> https://forum.minetest.net/vi​ewtopic.php?p=156409#p156409
16:32 kaeza I noticed there's no documentation about the "long" version of `drop` in lua_api.txt
16:32 indriApollo joined #minetest-dev
16:32 kaeza https://github.com/kaeza/minetest/commit/​c68c83a5f8398ed8168bc73a7b6848a9f429e804
16:36 indriApollo joined #minetest-dev
16:40 kaeza eh this one instead: https://github.com/kaeza/minetest/commit/​44b10eb2cb6452b633f20e6e804f62ef702987a9
16:44 BlockMen kaeza,  s/same node/node itself   ?
16:44 diemartin joined #minetest-dev
16:46 sfan5 BlockMen: do you manually add minetest_game to your builds?
16:47 BlockMen yes
16:47 zat joined #minetest-dev
16:47 sfan5 BlockMen: I just use the default install scripts
16:47 diemartin BlockMen, maybe
16:47 sfan5 that means jungles are broken for everyone that uses "make install"
16:47 BlockMen seems it misses something then?
16:48 sfan5 apparently
16:49 RealBadAngel string of the drop cannot hold the ammount?
16:50 RealBadAngel im not sure but i think it was able to do so
16:51 PilzAdam sfan5, BlockMen, https://github.com/minetest/minetest​/blob/master/CMakeLists.txt#L150-153
16:53 sfan5 who had that stupid idea?
16:53 sfan5 just install the whole game folder
16:55 diemartin RealBadAngel, looks like nope
16:55 RealBadAngel <George> Hi everyone I'm having trouble with mods not showing in games, hjow do I get then working?
16:56 RealBadAngel disabling mods by default is not a good idea
16:59 Hunterz joined #minetest-dev
17:00 PilzAdam RealBadAngel, enabling mods by default isn't either; it would apply to already created worlds without the user noticing it
17:00 PilzAdam a (IMO) good solution would be to show the "Configure" dialog in (or after) world creation
17:00 indriApollo left #minetest-dev
17:00 NakedFury joined #minetest-dev
17:00 BlockMen or enabled mods by default for new worlds
17:00 BlockMen *only
17:01 PilzAdam BlockMen, if you do that then please add a setting to disable that behavior; I have tons of mods and don't want to have them all enabled by default
17:02 BlockMen PilzAdam, it was just a suggestion... ;)
17:03 PilzAdam world creation could be "1) Choose a name, 2) Choose a game, 3) Choose a seed and mapgen (maybe games should be able to skip that step?), 4) Choose mods"
17:07 BlockMen ^ should games also be albe to skip that aswell?
17:07 BlockMen (4)
17:08 PilzAdam maybe mods could hint which games they work well on; and then theres a checkbox in 4) that says "Show mods that may not work with this game"
17:10 BlockMen no, not another cb
17:10 BlockMen the configuration menu is already messy
17:10 PilzAdam like "suggested_games = minetest_game, carbone" or "suggested_game = *all" in game.conf
17:12 PilzAdam BlockMen, redesign it then
17:12 BlockMen lel
17:13 PilzAdam don't limit your ideas to the current state ;-)
17:14 BlockMen im just try to limit making things worse ;)
17:16 PilzAdam there are only a few things that are worse than the current menu
17:19 BlockMen PilzAdam, everything can get worse
17:20 BlockMen but back to serious dev talk
17:20 BlockMen comments on https://github.com/minetest/minetest/pull/1690?
17:22 PilzAdam 1) why is it builtin? 2) I would prefer a /die with an extra priv
17:23 BlockMen why an extra priv for dying?
17:24 BlockMen you can also jump 3 times from a high place (in most cases)
17:24 PilzAdam imagine a grief_penalty mod that puts you in a 1x2x1 prison
17:26 BlockMen ok, but then the interact priv should be enough
17:35 sol_invictus joined #minetest-dev
17:52 Miner_48er joined #minetest-dev
17:57 BlockMen is that ok? https://github.com/BlockMen/minetest/commi​t/835f2fa84a9c62ef28e5b5969ccad44c2f898433
18:08 GrimKriegor joined #minetest-dev
18:51 Megaf_ PilzAdam: we have a jail in my server
18:51 Megaf_ made out of unbreakable nodes
18:52 Megaf_ and even if the user manages to escape it will teleport it back
18:52 Megaf_ PilzAdam: maybe this grief mod could use something like it
18:59 diemartin joined #minetest-dev
19:35 BlockMen left #minetest-dev
19:35 alexxs joined #minetest-dev
19:53 Amaz joined #minetest-dev
20:06 shadowzone joined #minetest-dev
20:20 EvergreenTree joined #minetest-dev
20:20 hmmmm http://www.wired.com/2014/10/code-p​ublished-for-unfixable-usb-attack/
20:27 ShadowNinja diemartin: The -- <...> should be ommited, simce it's not in a list and there's nothing obvious to go there.  Also, lua_api.txt isn't meant to be valid Lua, so you can remove the block comment notation.  Otherwise seems good.
20:29 diemartin ShadowNinja, https://github.com/kaeza/minetest/commit/​69476475a419777f9f6ed267ab4f4b553b6e0f44
20:30 diemartin actually, umm
20:30 ShadowNinja diemartin: Missing comma after items.items list, and spaces in {}.  I can fix that locally though.  hmmmm: Any comments?
20:31 hmmmm huh?  I'm confused
20:32 shadowzone So did the vote to ban/quiet prolloer go or not?
20:34 diemartin ShadowNinja, fixed https://github.com/kaeza/minetest/commit/​b8f9a98c61fe9c95daf333dbc3011ccfcccc14f5 (forgot a block)
20:34 shadowzone *proller
20:34 ShadowNinja shadowzone: Nothing's been done yet.  I'll check #minetest and notify ops here though.
20:34 shadowzone Okay.
20:35 ShadowNinja hmmmm: kaeza's patch ^.
20:35 Megaf_ ShadowNinja: the error message is as helpful as 99,9% of all minetest error messages!
20:35 Megaf_ https://github.com/minetest/minetes​t/issues/1697#issuecomment-57692540
20:35 Megaf_ I a system admin, not a bloody developer
20:35 Megaf_ system admins should be given meaning full error messages
20:36 Megaf_ not error messages that even developers can understand themselves
20:36 Megaf_ That's the first thing that should change in minetest
20:36 Megaf_ s/meaning full/meaningful
20:37 ShadowNinja Megaf_: I know that it's an error with memory management, but I need valgrind or GDB for details.
20:39 hmmmm ah
20:40 hmmmm no, no comments shadowninja
20:40 hmmmm LGTM
20:46 ShadowNinja diemartin: Actually it's already mentioned, see line 2424.  I'll replace it though.
20:47 Megaf_ Is this a bug on my mod, on pipeworks or a minetest bug? https://github.com/Megaf/more_chests/commi​t/4aa321957497512ea012736aa2b594818d328dd4
20:47 diemartin ShadowNinja, oh, didn't notice that one
20:47 ShadowNinja diemartin: (I'm fixing it locally)
20:47 diemartin ok, thanks
20:50 ShadowNinja kaeza: https://github.com/minetest/minetest/commi​t/5baf379f40537060a1d5111905c7702d4563ee5c
20:51 kaeza and even with a better commit message :D
20:52 Megaf_ VanessaE: https://github.com/VanessaE/pipewor​ks/issues/74#issuecomment-57704547
20:52 rmilan joined #minetest-dev
20:53 srifqi joined #minetest-dev
20:53 srifqi left #minetest-dev
20:56 shadowzone joined #minetest-dev
20:56 shadowzone joined #minetest-dev
20:57 kaeza joined #minetest-dev
21:07 proller joined #minetest-dev
21:31 Nerotic joined #minetest-dev
21:31 Nerotic left #minetest-dev
21:35 casimir joined #minetest-dev
22:07 chchjesus joined #minetest-dev
22:07 eeew joined #minetest-dev
22:27 shadowzone joined #minetest-dev
22:27 shadowzone joined #minetest-dev
22:29 rickmcfarley joined #minetest-dev
22:33 paramat joined #minetest-dev
23:06 ShadowNinja Calling those modpack-with-menu-things "Games" seems to be causing real confusion: https://forum.minetest.net/v​iewtopic.php?f=3&amp;t=10254
23:07 ShadowNinja "Subgames" seems to be the best suggestion so far.
23:14 paramat Who cares about confused newbies? they'll soon learn, too much spoonfeeding ... the engine is not a game, subgames are not sub-, they are games =)
23:18 rickmcfarley joined #minetest-dev
23:19 rickmcfarley I've had to explain the difference to a few new comers
23:20 rickmcfarley I wish there was a launcher to view/add games
23:21 rickmcfarley I was thinking about making a simple python script, but figured there was already such a thing
23:25 AnimalMachine joined #minetest-dev
23:35 paramat left #minetest-dev
23:48 rickmcfarley joined #minetest-dev
23:52 Weedy joined #minetest-dev
23:55 SmugLeaf joined #minetest-dev

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