Minetest logo

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

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

All times shown according to UTC.

Time Nick Message
00:00 RealBadAngel i would ratther wait for jin_xi's code
00:01 RealBadAngel native particles are faster
00:01 Taoki joined #minetest-dev
00:02 kahrl it's not about performance *at all*
00:02 RealBadAngel well, i can make them dependent on the crack setting
00:03 kahrl I think that would satisfy everyone
00:03 RealBadAngel what would satisfy everone would be nice menu
00:04 RealBadAngel with slider for it, as well as for other stuff
00:04 kahrl I don't want to sit in the menu tweaking 100 settings before I can enjoy a game
00:05 RealBadAngel but you want to edit txt file
00:05 kahrl tweaking 1 setting should work out of the box and not require tweaking lots more before I looks ok
00:05 kahrl else you might as well be told to code the engine yourself
00:05 RealBadAngel or you just want to block the change making it extremaly unaviable for newbies...]
00:06 kahrl s/I/it
00:06 RealBadAngel no adroid user will be able to use the change
00:07 RealBadAngel our menu system is stuck, in dead end
00:07 RealBadAngel theres no place to add anything new
00:07 kahrl why is there no text editor on android anyway
00:07 RealBadAngel youre coder unlike the most players
00:08 RealBadAngel if you will continue to think that way we will loose players
00:08 RealBadAngel as we continously do
00:08 RealBadAngel mt is not newbie friendly
00:08 RealBadAngel never was
00:09 kahrl what do you think a player prefers? "uncheck the 'crack animation' checkbox" or "uncheck the 'crack animation' checkbox and, oh I almost forgot, you also need to increase the number of punching particles on settings page #42"
00:09 RealBadAngel i think they want nice gfx effect
00:09 RealBadAngel and fast
00:10 RealBadAngel but our way is to make settings
00:10 RealBadAngel which ends with most of the features never used
00:10 RealBadAngel just by hardcore freaks
00:11 RealBadAngel have you ever seen comments on android for our clones?
00:11 RealBadAngel i saw some in polish
00:12 kahrl I don't care about the clones
00:13 RealBadAngel but they were about basic things
00:13 RealBadAngel our game is hardcore
00:13 RealBadAngel even the mods cannot be easily installed
00:13 RealBadAngel see names of folders issue
00:14 RealBadAngel if i see such question i dont answer anymore, im tired of this shit
00:14 kahrl how is that related to anything?
00:15 RealBadAngel its the same
00:15 RealBadAngel its about the gameplay
00:15 kahrl and I think the authors of the clones just cared about quick buck and not ease of use of anything
00:16 RealBadAngel they just showed us to others
00:16 RealBadAngel and folks commented our errors not the the fork makers
00:17 kahrl remember that this is primarily a PC game
00:17 RealBadAngel how long the naming folders issue has to be here to solve it?
00:17 RealBadAngel why the fuck name of the folder has to mean anything?
00:17 kahrl it's not our fault if someone does a bad porting to mobile job (and remember that was before our android port)
00:17 RealBadAngel answer above
00:18 RealBadAngel how many hundreds of times have you saw error reports because of that?
00:18 kahrl I'm tired of the mod name discussion so I won't answer
00:18 kahrl it leads to nowhere
00:19 RealBadAngel i pointed the solution
00:19 RealBadAngel file with modname
00:19 RealBadAngel but nobody cares
00:19 RealBadAngel better to answer (or not) the error reports
00:20 RealBadAngel this is one of the features i requested from other devs
00:20 RealBadAngel but everybody has that deep in the ...
00:21 RealBadAngel same for database key storage
00:24 kahrl what about database key storage?
00:24 mos_basik_ joined #minetest-dev
00:26 RealBadAngel store lua tables in a database
00:26 paramat MT is a little newbie-unfriendly, but that's good, it filters out those with little commitment, and the stupid people
00:26 RealBadAngel like per player mod settings ec
00:26 RealBadAngel paramat, atm it filters all the players out there
00:27 jin_xi paramat: its a game suitable for the very young so it needs to be newbie friendly imo
00:27 RealBadAngel not everybody is an IT expert
00:27 paramat hm, well good point
00:27 kahrl RealBadAngel: not a topic I followed, link to issue?
00:28 RealBadAngel kahrl, after a several months? selat already made that for freeminer
00:28 kahrl cool
00:28 paramat RBA perhaps you could keep 2 forks of MT, one being exactly how you want it, for your own use, second fork is where you compromise a little more, so we can benefit from your good work
00:29 RealBadAngel kahrl, not cool because mt cant have it
00:29 RealBadAngel same for circuits
00:29 RealBadAngel another shame for us
00:30 RealBadAngel if you dont know, redstone was one of the pillars of mc success
00:30 RealBadAngel we refused to have stable and fast circuits simulator
00:31 paramat in my mods i often state 'change mod folder name to ...' to help people
00:31 RealBadAngel we just follow the same old rule, shoot own foot each time we can aim at something
00:31 RealBadAngel just to scare everybody around
00:32 jin_xi hm, im still thinking about that texture leak. why do we keep all textures when they are recombined?
00:32 RealBadAngel paramat, thats not a solution, by any meaning
00:34 paramat it's obvious you want to make big, controversial changes to MT, and have a very different AV style, so to contribute you may need to compromise a little
00:34 kahrl I never really looked at the circuit thing either because it's outside my area, but I figured there wasn't much interest because mesecons exists
00:35 RealBadAngel kahrl, so the others took it
00:35 kahrl and that's cool
00:35 RealBadAngel no
00:35 RealBadAngel that means that mesecons are not widely used
00:35 RealBadAngel becaue theyre server killer
00:36 RealBadAngel mods are too slow for them to be widely used in multiplayer enviroment
00:36 kahrl and how is a pull request more "stable" than a widely tested mod?
00:36 RealBadAngel so mesecons will be never as cool for new players
00:37 kahrl at the time I looked there was no reports of any tests
00:37 RealBadAngel because servers wont use it just
00:37 RealBadAngel for sake of speed
00:37 RealBadAngel kahrl, even minetest classic has circuits, where have you been last few months?
00:38 Taoki joined #minetest-dev
00:38 kaeza joined #minetest-dev
00:38 RealBadAngel we are killing anything new that tries to come
00:41 rickmcfarley joined #minetest-dev
00:41 kahrl again, not my area, ask the other devs
00:41 paramat RBA, from following IRC and the forum, you are possibly our most popular dev, MT community (and myself) are very keen to see your work get merged, so please compromise a little? =) *pweez*
00:42 kahrl RealBadAngel: all I can say from quickly looking at the PR is that the textures look much worse compared to mesecons
00:43 RealBadAngel 09:42:49 < celeron55> i'm surprised how well that is done
00:43 RealBadAngel textures are secondary
00:43 kahrl I guess the textures gave the other devs a bad impression
00:43 RealBadAngel and absolutely had no meaning at all
00:44 RealBadAngel code was important
00:45 kahrl first impression is always important
00:46 RealBadAngel not every coder can make good gfx for his code
00:46 RealBadAngel his work should be completed with further changes
00:46 RealBadAngel like the visuals
00:46 kahrl so since this has been used for a while, I'm sure someone made some textures?
00:47 RealBadAngel in mt?
00:47 RealBadAngel i dont think so
00:47 kahrl no, in classic or fm
00:47 RealBadAngel we forced selat to move to fm
00:47 RealBadAngel another coder lost
00:47 proller cool
00:48 RealBadAngel how many coder we need to lost to understand something?
00:48 RealBadAngel all of them?
00:49 kahrl I'm the only dev reading this and I wasn't the one who blocked those, so what do you intend my telling me this?
00:49 kahrl s/my/by
00:51 RealBadAngel others can propably read that too
00:52 RealBadAngel i can see hmmmm sfan5 celeron55 ShadowNinja online
00:52 RealBadAngel and proller and darkrose laughing
00:53 proller HA-HA-HA --^
00:53 kahrl those didn't block the circuit PR (not sure about the other PRs)
00:54 RealBadAngel kahrl, circuits were just an example (important anyway)
00:54 RealBadAngel nothing of value gets merged
00:55 RealBadAngel rarely some slight fixes or secondary features
00:55 kahrl and what is your constructive idea to change that?
00:55 RealBadAngel which are copies of already existing ones
00:56 RealBadAngel stop treating dev as stable
00:56 kahrl and I see quite a lot of major stuff on http://dev.minetest.net/Changelog
00:56 RealBadAngel this is one of the reasons new code is not merged
00:56 kahrl although I agree it has slowed down lately
00:57 kahrl RealBadAngel: users are treating dev as stable
00:57 RealBadAngel blockers want to be it perfect
00:57 kahrl so we essentially have to
00:57 RealBadAngel while it should be tested and improved over time
00:57 RealBadAngel no we dont have to
00:57 RealBadAngel stable should be used as its called so
00:58 RealBadAngel dev should be devs branch
00:58 kahrl not possible with our release schedule
00:58 RealBadAngel whoever wants to help us test and improve the code then he should touch it
00:58 kahrl btw, a lot of the debugging I've done lately was through git bisect, I will stop doing that if every other commit fails to start
01:01 RealBadAngel kahrl, i was testing highlight for weeks before i pushed it
01:02 RealBadAngel then vanessa found two issues in an hour
01:02 Zeno` joined #minetest-dev
01:02 RealBadAngel theres no other way i mean
01:02 RealBadAngel code has to be made public in dev
01:02 Zeno` Morning.
01:03 RealBadAngel then more folks will test it and help improve
01:03 kahrl well, we could probably be a bit less perfectionist, but it has to be carefully balanced lest we end up like fm
01:03 RealBadAngel thats the fucking meaning of development branch
01:03 Zeno` 1729 I would call a bug fix rather than maintenance because using unititialised variables is undefined behaviour ;)
01:03 RealBadAngel kahrl, ofc
01:03 Zeno` err... bad timing
01:03 RealBadAngel merging anything could be wrong
01:04 RealBadAngel but merging things going in good direction, and help it is the other side
01:05 RealBadAngel selat was propably unable to make circuits complete alone
01:05 RealBadAngel he needed good drawtype, textures etc
01:05 RealBadAngel who gave him a hand?
01:05 RealBadAngel not us
01:06 paramat the absence of the AV manager while RBA is around is a problem, i would like to see 2 other core devs get together to discuss and merge RBA's work
01:07 RealBadAngel i will be able to make a PR within a few days
01:07 RealBadAngel meshnodes are the most complex change to mt i ever tried
01:08 RealBadAngel mapblock_mesh code is unbeliveably short
01:08 RealBadAngel but thx to the supporting code
01:08 RealBadAngel that change touches over 20 source files
01:09 kahrl paramat: that would be great, but I don't see it happening
01:09 RealBadAngel if i will rebase it i will do it just once
01:10 RealBadAngel i wont be waiting months sitting here and asking
01:10 paramat i will fight to keep you with us RBA, and for your work to be merged, as long as artistic AV changes are kept separate (because they will be controversial due to your different AV style)
01:10 kahrl most devs are rarely here and typically miss each other by a few hours
01:10 paramat hm indeed
01:10 kahrl which can be done productively in an asynchronous medium like IRC, but it isn't done here
01:11 kahrl (I don't blame them, nobody is paid)
01:12 RealBadAngel anyway, when im done with meshnodes, next step is to move highlighting and cracks away from mapblock_mesh at all
01:12 RealBadAngel its a bottleneck
01:12 RealBadAngel next is to turn nodeboxes into meshes on the fly
01:13 RealBadAngel next the shaders update (thx to having meshes i can convert them all into meshes with tangents)
01:14 RealBadAngel when i will have tangent space aviable i can turn on the lights
01:14 RealBadAngel as you can see each piece of what im doing is on purpose and requiring previous work
01:15 RealBadAngel if you still want to discuss if 3 particles are worth blockin my code, feel free to do so
01:16 kahrl I suggested a way to change it
01:16 RealBadAngel thats not the point
01:17 RealBadAngel if i dont have enough power to force 3 particles to go in, what the fuck i can do at all?
01:17 RealBadAngel and whats the point of the efforts in the first place?
01:17 kahrl nobody needs to be able to "force" anything
01:18 RealBadAngel im the only one who is coding AV
01:18 RealBadAngel and im in charge of nothing :P
01:18 kahrl working in a team means you have to accept compromises sometimes
01:19 RealBadAngel thats why we have settings
01:19 RealBadAngel so most of the players wont see the changes
01:19 kahrl I already talked about this in #1710, but if you think you need to be able to force anything I will retract my agreement to make you AV maintainer
01:19 RealBadAngel and c55 good sleep is safe
01:21 RealBadAngel kahrl, you know im not forcing anything
01:21 kahrl you said so yourself
01:21 RealBadAngel remember highlighting? how it evolved?
01:22 RealBadAngel i was listening even to paramat with his extremaly voxel approach
01:22 RealBadAngel and tried to satisfy everybody out there
01:23 kahrl yeah, because that's what it took to get the PR merged
01:23 kahrl if you could just merge it nilly willy you probably wouldn't have listened to everything
01:24 RealBadAngel i know when im doing something to be tuned
01:24 kahrl see: 1705
01:24 RealBadAngel but damn 3 particles?
01:24 RealBadAngel or fuckin name for a bit?
01:24 kahrl you don't care about getting it fixed because your code is already merged
01:25 RealBadAngel ofc i do care
01:25 RealBadAngel thats why code from shaders was removed
01:25 RealBadAngel thats one of the steps to turn on the lights
01:26 RealBadAngel cpu and gpu doing the same all the time was just ridiculous
01:26 RealBadAngel and PA knows that
01:26 kahrl they weren't doing the same
01:26 RealBadAngel yes they were, changing vertex colors
01:27 kahrl doing it only on the gpu was the point of the old code which you call faulty
01:27 RealBadAngel we cant do it only on GPU side
01:27 paramat i will support your meshnodes pull RBA. the reason i chatted here tonight is because i saw you were close to a pull request, and i want to see it merged. fromour chat it looks good to me so far
01:28 RealBadAngel not all will have shaders
01:28 paramat i wanted to check for things that might get it blocked
01:28 RealBadAngel paramat, that one is pure tool as you may call it
01:29 kahrl RealBadAngel: ok, let me assume as devil's advocate that both cpu and gpu did it. You removed the gpu part, so by elementary logic it should still be working because the cpu still does it. Contradiction, so only the GPU did it
01:29 RealBadAngel theres no AV change unless you do it
01:29 paramat yes cool
01:29 RealBadAngel kahrl, no, its done by cpu all the time
01:29 kahrl RealBadAngel: I just proved mathematically that it isn't
01:29 RealBadAngel but threading shows how bad it is
01:30 RealBadAngel but it is very same
01:30 RealBadAngel shaders were just not slowed down by threads
01:30 RealBadAngel but thats not the point
01:31 kahrl it has nothing to do with threading
01:31 kahrl meshes are all animated in the same thread
01:31 RealBadAngel so maybe something other is slowing them down
01:32 kahrl but not every mesh in animated every frame (try changing isAnimationForced() to return true and you'll see why)
01:32 RealBadAngel anyway, that system is sick
01:33 kahrl how so?
01:33 RealBadAngel its server side
01:33 kahrl wat
01:33 RealBadAngel lighting is calculated server side
01:33 RealBadAngel then passed to client
01:34 kahrl oh, I thought you meant the animation system
01:34 RealBadAngel its done here
01:34 RealBadAngel by changing vertex colors of all the meshes visible
01:34 kahrl I don't want to talk about the lighting system again
01:34 RealBadAngel you started
01:34 kahrl well, not now at least
01:34 kahrl no
01:35 RealBadAngel with 1705
01:35 kahrl that's more an animation system issue
01:36 kahrl I didn't look at the code so it might be obvious, but why can't finalColorBlend stay in the shader until hardware lighting is implemented?
01:37 RealBadAngel i can make it in again
01:37 RealBadAngel but if i know its obsolete?
01:38 kahrl do it
01:38 kahrl otherwise you're blocking a release until you finish hardware lighting and we agree to merge it
01:38 RealBadAngel then face another discussion why i removed it... dead circle
01:39 kahrl well, I assume it won't be needed anymore with hardware lighting
01:39 kahrl so that discussion will be over quick
01:40 RealBadAngel nodeboxes are dead already also
01:40 kahrl nope
01:40 RealBadAngel yup
01:40 kahrl they aren't
01:40 RealBadAngel who will want to make game slower?
01:40 RealBadAngel or with worse effects?
01:40 kahrl they don't make the game slower since you convert them to meshes
01:41 RealBadAngel if r2d2 is faster than stair, think how mesh stair will be faster
01:41 RealBadAngel i will do that, but with a price
01:41 kahrl DLC?
01:41 RealBadAngel lol
01:41 RealBadAngel no
01:41 RealBadAngel waiting time
01:42 mos_basik_ joined #minetest-dev
01:42 RealBadAngel they will have to be processed
01:42 kahrl how long does it take per mesh?
01:42 RealBadAngel meshes are pretty quick
01:43 kahrl I mean, how long does converting one node's nodebox to mesh take
01:43 RealBadAngel slower nodebox code will have to be called for each node to get mesh buffer
01:43 RealBadAngel then the buffer has to be cloned 23 times
01:43 RealBadAngel of if its facedire
01:43 RealBadAngel +d
01:44 RealBadAngel http://pastebin.com/tYjiczUn
01:45 RealBadAngel instead of getMesh from file mapblock mesh method fillSingle has to be called
01:45 RealBadAngel for each nodebox out there
01:45 kahrl uh
01:46 kahrl why not just directly call the code that is in case NDT_NODEBOX right now
01:47 RealBadAngel this is not runtime code
01:47 kahrl ?
01:47 RealBadAngel this is nodedef.cpp
01:47 kahrl I don't understand the problem
01:48 RealBadAngel i can call it, theres already a method rdy to get meshbuffer for single node
01:49 RealBadAngel which im going to use to convert nodeboxes into meshes
01:49 RealBadAngel i just said it gonna take time
01:49 kahrl RealBadAngel: it's incredibly wasteful if you don't need the extra things MapBlockMesh does
01:50 RealBadAngel mapblock_mesh does so many unnecesary things thats just incredible ;)
01:51 RealBadAngel and most important is that it is doing them all over
01:51 RealBadAngel which slow down the client as hell
01:51 RealBadAngel previously it was even able to call for files
01:52 kahrl if starting a game with a decent amount of nodeboxes takes a long time because of a dumb thing like that, I will block your changes :P
01:53 RealBadAngel if you will block things that can raise fps by half or something like that....
01:53 RealBadAngel feel free to do so :P
01:53 RealBadAngel shot the feet, go on :)
01:54 kahrl sure, I will block it
01:54 kahrl but I told you how to fix it
01:54 RealBadAngel wait, where?
01:54 RealBadAngel <kahrl> why not just directly call the code that is in case NDT_NODEBOX right now
01:54 RealBadAngel here?
01:54 kahrl yes
01:54 RealBadAngel uh
01:55 RealBadAngel read what ndt nodebox does first
01:55 RealBadAngel then come back again
01:55 kahrl what about it?
01:55 RealBadAngel it reads the boxes
01:55 RealBadAngel rotates them
01:55 RealBadAngel fixes them
01:55 Zeno` I think most people would prefer slower load times if FPS was increased
01:55 RealBadAngel makes cuboids
01:56 RealBadAngel maps textures
01:56 RealBadAngel before rotates the textures
01:56 RealBadAngel then appends the collector
01:56 kahrl RealBadAngel: that's exactly what's needed to make a mesh, isn't it?
01:56 RealBadAngel mesh does the last step only
01:57 RealBadAngel but all the steps have to be done at least once
01:57 paramat i would happily wait a couple of seconds on startup for faster nodeboxes done by meshnodes
01:57 RealBadAngel at startup
01:58 RealBadAngel i will have to call mapblock mesh for each and everynodebox defined in the game to get its meshbuffer
01:58 RealBadAngel then i can treat is as meshnode
01:58 kahrl no
01:58 kahrl just call the damn NDT_NODEBOX code (after moving it to a new function of course)
01:58 kahrl no need to do the same thing in a roundabout way
01:59 RealBadAngel kahrl, theres a method for that
01:59 RealBadAngel in mapblock mesh
01:59 kahrl yeah but it's inefficient
01:59 RealBadAngel propably yes, but it is here
01:59 kahrl you're free to optimize that function instead
02:00 RealBadAngel i can do that
02:00 RealBadAngel i havent started with that
02:00 kahrl I fear that after you get your pulls merged, you don't have incentive to do so, because "node boxes are a thing of the past"
02:01 RealBadAngel whats left for me now is copying bounding boxes to selection boxes, and make inventory images for meshes
02:01 RealBadAngel all the other stuff is done
02:02 paramat dreambuilder with 1000s of nodeboxes might be slow though, but then that'a probably slow at startup anyway
02:05 RealBadAngel dreambuilder is unplayable atm
02:05 RealBadAngel join spawn area
02:05 paramat lol
02:05 RealBadAngel you will get a few fps
02:05 RealBadAngel courtessy of nodeboxes (tm)
02:06 RealBadAngel also the signs, but thats other story ;)
02:06 paramat fast in game justifies slower startup
02:07 RealBadAngel im getting 3-5 fps on VE dreambuilder server
02:07 Zeno` I just want to get more than 20fps on dual gtx-780 and i7
02:07 Zeno` so I can play the silly game hehe
02:08 RealBadAngel theres no other way to make it fast than coding it fast
02:09 RealBadAngel if nodeboxes are a bottleneck, we shall consider throwing them away
02:09 RealBadAngel but wait
02:09 RealBadAngel modders will do that for us :P
02:10 RealBadAngel nobody will want to ride a bike with a shine new ferrari on the parking lot :P
02:10 kahrl some of them probably, but a lot of them will stay nodeboxes
02:11 kahrl they are just way more convenient to write, modify and generate programmatically
02:11 RealBadAngel who told you that?
02:11 kahrl this opinion has been voiced by various people in this channel
02:11 jin_xi i think implementation is one thing format another. nodeboxes is a fine format.
02:11 kahrl just grep the logs
02:12 jin_xi if it can be handled more efficiently, then do that by all means
02:12 kahrl RealBadAngel: remember that you yourself had to ask someone else to make a full node mesh in blender for you
02:12 RealBadAngel it was two days ago
02:12 kahrl while as nodebox it's 3 simple lines
02:12 RealBadAngel now im able to make such things:
02:13 kahrl RealBadAngel: yeah, you, but not every modder coded the mesh drawtype
02:13 kahrl and has become a blender wizard
02:13 RealBadAngel http://i.imgur.com/ECELl3Y.png
02:13 jin_xi well, i think its nice to open the world of minetest to meshes but its not mesh against nodebox
02:14 RealBadAngel nodebox is a way to describe the mesh
02:14 RealBadAngel faulty and slow
02:14 RealBadAngel and with so many limitations thats incredible
02:14 kahrl RealBadAngel: I really hate that whenever you make something new, you think that whatever was there before is completely inferior and must be eradicated
02:15 RealBadAngel maybe
02:15 RealBadAngel but timings are cruel
02:15 RealBadAngel nodeboxes suck big time
02:16 RealBadAngel gtg, cya tommorow
02:17 Zeno` cya
02:19 kahrl cya
02:34 paramat left #minetest-dev
02:36 DuDraig joined #minetest-dev
03:12 rickmcfarley left #minetest-dev
03:28 mos_basik__ joined #minetest-dev
03:32 kaeza joined #minetest-dev
03:36 GrimKriegor joined #minetest-dev
03:48 Hunterz joined #minetest-dev
04:27 Miner_48er joined #minetest-dev
04:38 paramat joined #minetest-dev
04:42 paramat hmmmm, in mapgen.cpp, 'lightSpread()' seems to call itself 6 times, is this to create a huge exponential cascade of light-spreads? or perhaps i misunderstand ...
04:43 hmmmm it's how light is spread
04:43 hmmmm it just keeps recursively spreading itself in each direction until the light has completely decayed
04:45 paramat okay thanks, it was just a bit shocking to see
04:58 hmmmm hrmmm
04:59 hmmmm is it just me or weren't decorations used in minetest_game
04:59 hmmmm and now they aren't...
05:00 paramat dunno, they're used in mgv6 too?
05:01 Miner_48er joined #minetest-dev
05:08 hmmmm i am really confused
05:09 hmmmm either I'm crazy, or at one point in minetest_game, grass used the decoration API for placement
05:09 hmmmm same with cacti
05:23 hmmmm this is pretty hilarious
05:24 hmmmm so it seems like I was mistaken the whole while, minetest_game never used any decorations, and my code is broken because stone in the shape of trees surrounded by fire is sprouting up everywhere, all of a sudden
05:25 Miner_48er hmmm that would be an interesting mod
05:33 sol_invictus joined #minetest-dev
05:38 paramat fire ... the ultimate error message of voxelmanip
05:40 Hunterz joined #minetest-dev
05:47 mos_basik__ joined #minetest-dev
05:48 mos_basik__ joined #minetest-dev
05:50 celeron55 RealBadAngel: how are you batching meshnodes?
05:50 celeron55 does it increase context switches compared to nodeboxes
05:50 celeron55 or draw calls
06:10 paramat woo! just pushed your mgv5 branch commits into latest git and have mgv5 running
06:12 paramat just one error message during compiling: In function ‘ll_sym’: /home/***/minetestd1cmgv5/s​rc/lua/src/loadlib.c:76:21: warning: ISO C forbids conversion of object pointer to function pointer type [-Wpedantic]   lua_CFunction f = (lua_CFunction)dlsym(lib, sym);
06:12 paramat left #minetest-dev
06:14 Hunterz anyone can help me wit this error? http://pastebin.com/bD3wYd3F
06:17 hmmmm seems like it's trying to use lua as luajit
06:18 Hunterz ah, thanks
06:18 hmmmm should manually override the luajit library to be /usr/lib64/libluajit-5.1.so
06:18 Hunterz it works now, many thanks for help
06:27 hmmmm paramat, that's a warning, not an error
06:28 hmmmm the problem is that gcc does not provide a POSIX mode with its -std=ansi or what ever
06:38 asl joined #minetest-dev
06:45 Zeno` you can get rid of the warning if you really want to
06:46 Zeno` (without removing -pedantic of course)
06:48 Zeno` http://codepad.org/Ny4pnCy7
06:50 hmmmm he can't modify that because it's a 3rd party lib
06:50 Zeno` yeah, which is why I said "if you want to" :)
06:50 Zeno` But what do you mean can't modify the 3rd party lib?
06:51 Zeno` It's right there in the src tree
06:52 Zeno` I wouldn't but it doesn't mean you can't
06:52 celeron55 well gcc has this functionality that works like "#pragma GCC diagnostic push" which can be used to ignore warnings from includes 8))
06:53 Zeno` if you look at the man page for dlsym the way I did it is the way that POSIX recommends it be done, so Lua needs fixing upstream
06:53 Zeno` or there is the pragma, yes
06:53 celeron55 go complain to the lua devs then, we have nothing to do with thatt
06:53 celeron55 -t
06:53 Zeno` I didn't bring it up
06:54 celeron55 dunno about lua's policies about this kind of stuff though
06:54 Zeno` Personally I just ignore the warning
07:05 shmancelot joined #minetest-dev
07:07 RealBadAngel celeron55, meshes are collected with a slightly modified append method. only difference to the original one is that it applies there color and pos to the vertices
07:09 RealBadAngel http://pastebin.com/wQMknFEH
07:11 RealBadAngel at this point theres no difference its meshnode or any other drawtype
07:14 celeron55 so they are just added to the mesh like anything else
07:16 celeron55 it desperately needs texture atlases but that's been the case ever since they were removed so whatever
07:18 RealBadAngel yes, theyre treated the same
07:19 RealBadAngel ofc this way only static ones can be added
07:20 RealBadAngel for animated ones i will have to code something in between mesh collector and entities
07:20 RealBadAngel will solve that when removing highlighting and cracks from mapblock_mesh
07:21 RealBadAngel both are causing unnecesary mesh updates
07:25 ImQ009 joined #minetest-dev
07:27 RealBadAngel texture atlases were a problem when we needed rotations for textures
07:27 RealBadAngel that was a main reason to remove it
07:50 Krock joined #minetest-dev
08:24 darkrose joined #minetest-dev
08:29 FR^2 joined #minetest-dev
08:46 celeron55 it must be possible to handle it
08:46 celeron55 they would increase performance a huge amount
08:47 celeron55 removing them was just laziness
08:53 * Krock is interested. opens log
09:26 Amaz joined #minetest-dev
09:39 Amaz joined #minetest-dev
09:52 Krock Wait. can I disable the stairs on the sky?
10:02 elinvention joined #minetest-dev
10:05 proller joined #minetest-dev
10:16 PenguinDad joined #minetest-dev
10:19 kilbith joined #minetest-dev
10:33 proller joined #minetest-dev
10:43 jin_xi joined #minetest-dev
11:18 Zeno` Judging by the rate of commits this seems more like a 1.x.x version than a 0.x.x version
11:36 GrimKriegor joined #minetest-dev
11:36 Zeno` Even bug fixes aren't being merged :/
11:39 proller nobody cares
12:02 Amaz proller, I do.
12:56 shadowzone joined #minetest-dev
13:08 Krock joined #minetest-dev
13:12 NakedFury joined #minetest-dev
13:15 petrusd987 joined #minetest-dev
13:33 proller joined #minetest-dev
13:35 kilbith joined #minetest-dev
13:51 prozacgod joined #minetest-dev
13:56 proller joined #minetest-dev
13:57 hmmmm joined #minetest-dev
14:07 Hunterz joined #minetest-dev
14:07 rmilan joined #minetest-dev
14:24 proller joined #minetest-dev
14:30 Krock ppl here either gonna hate me or like..
14:31 Krock #1731
14:36 eeew joined #minetest-dev
14:45 Zeno` if(net_proto_version >= LATEST_PROTOCOL_VERSION - 1 &&     <-- is that tested?
14:45 Zeno` if should have a space before the ( as well
14:45 Krock Zeno`, everything's untested
14:46 Zeno` The condition seems back to front to me
14:47 Zeno` just glancing at it I would have thought it should be   if (net_proto_version < LATEST_PROTOCOL_VERSION - 1 && net_proto_version > LATEST_PROTOCOL_VERSION + 1) {
14:47 Krock Should I make a pull request to only convert all "if(" to "if ("?!
14:48 Krock damn I did it the worong way there
14:49 Krock mhm the callenge is now to edit that specific commit
14:51 Zeno` just amend last commit and force push
14:51 Krock net_proto_version < LATEST_PROTOCOL_VERSION - 1 || net_proto_version > LATEST_PROTOCOL_VERSION + 1) { prevent from connecting }
14:54 Zeno` that seems right
14:54 Krock there we go.
14:55 Zeno` good luck :p
14:55 Krock damn. need to rewrite the doc again
15:01 * Krock gives Zeno` a cookie
16:12 Calinou joined #minetest-dev
16:18 zat joined #minetest-dev
16:35 gravgun joined #minetest-dev
16:55 kaeza joined #minetest-dev
17:08 HLuaBot joined #minetest-dev
17:18 Calinou joined #minetest-dev
17:37 Hunterz joined #minetest-dev
17:39 alexxs joined #minetest-dev
17:42 alexxss joined #minetest-dev
17:44 Du_Draig joined #minetest-dev
17:58 Megaf kahrl: sfan5: kilbith: It failed to compile a static bin. http://paste.debian.net/plain/125609
17:58 RiZom-91 joined #minetest-dev
17:59 sfan5 thats because you don't have the static libs for those
17:59 sfan5 Megaf: just compile a non-static binary, it probably requires raspbian to use anyway
18:18 FR^2 joined #minetest-dev
18:24 Sokomine commented on https://github.com/minetest/minetest/pull/1732  i really hope it will be merged!
18:27 Anchakor_ joined #minetest-dev
18:35 kilbith_ joined #minetest-dev
18:35 kilbith_ Megaf: you can solve that ?
18:35 Megaf nope
18:35 Megaf Im rebuilding it no static
18:35 kilbith_ like yesterday so ?
18:36 kilbith_ we'll have the DRI2 bug again...
18:42 paramat joined #minetest-dev
18:55 Sokomine i've got a problem with lua/luajit under windows. unlike the linux version, it does not seem to read the file entirely. does anyone have any clues regarding that? or knowledge of a more appropriate channel?
18:57 pitriss Sokomine: hmm what about different line ends on windows?
18:58 Sokomine pitriss: it's a binary file and read byte-wise
18:58 pitriss ahh then IDK
18:59 jin_xi joined #minetest-dev
18:59 Sokomine sample code is here: https://github.com/Sokomine/mg_vi​llages/blob/master/read_test.lua   it does read far less bytes under windows than under linux. only thing i could think of would be encountering an eof in the middle of the file?
19:01 kaeza Sokomine, open with "rb"
19:02 Sokomine kaeza: seems that did it! thank you :-)
19:08 diemartin joined #minetest-dev
19:26 zat joined #minetest-dev
19:27 DuDraig joined #minetest-dev
19:30 Miner_48er joined #minetest-dev
19:32 Amaz joined #minetest-dev
19:43 sapier joined #minetest-dev
19:55 SudoAptGetPlay joined #minetest-dev
19:57 SudoAptGetPlay Hi, I build the latest minetest from master branch on ubuntu 14.04 an hour ago and I'm getting plagued by messages like "Irrlicht log: Could not open file of texture: UV_rat.png" constantly. An idea ? The rats are displayed correctly by the way, but the game hangs a bit.
19:59 Amaz Which mod are you using?
20:00 SudoAptGetPlay Mobs mods
20:00 SudoAptGetPlay It's not local it's on a server actualy
20:01 Calinou SudoAptGetPlay, fixed in my mobs mod
20:01 sapier ignore it most likely its an unimportant error in the model
20:01 Calinou Simple Mobs includes a reference to an UV file
20:01 Calinou edit .x file, remove reference
20:01 Calinou fixed!
20:02 SudoAptGetPlay The error doesn't affect the drawing of the textures, but the game freezes for a couple of millisecs each time that error occurs. And I got like 10 errors a min
20:02 sapier millisecs?
20:02 sapier that's strange
20:02 SudoAptGetPlay The game hangs, I recieve lots of errors and then I can go on
20:03 sapier can you try disabling logging?
20:03 sapier and delete debug logfile
20:03 SudoAptGetPlay I launched the game in a console
20:03 sapier I guess it's a disk io issue
20:03 SudoAptGetPlay bad for me I guess
20:03 sapier how big is your debug.txt file?
20:04 SudoAptGetPlay nothing is logged in debug.txt, the logs are printed on my terminal
20:05 SudoAptGetPlay The time we spoke only :
20:05 SudoAptGetPlay Could not open file of texture: sheep.png
20:05 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 kilbith_ STOP
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 kilbith_ ^ sfan5
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:05 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:06 sapier SudoAptGetPlay: NEVER post a logfile to irc!!!!!!!!!!!!!!!!!
20:06 SudoAptGetPlay Irrlicht log: Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:06 SudoAptGetPlay Could not open file of texture: sheep.png
20:07 kilbith_ someone can kick ?
20:07 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:07 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:07 SudoAptGetPlay Could not open file of texture: UV_rat.png
20:07 SudoAptGetPlay don't worry that's all ^^
20:07 SudoAptGetPlay srry
20:07 SudoAptGetPlay It'sa short one
20:07 SudoAptGetPlay But yeah no spam ...
20:07 kilbith_ bravo hein...
20:07 kilbith_ utilise pastebin la prochaine fois
20:08 MinetestForFun joined #minetest-dev
20:30 shadowzone left #minetest-dev
20:39 shadowzone joined #minetest-dev
20:39 shadowzone joined #minetest-dev
20:56 mrtux joined #minetest-dev
21:27 mos_basik joined #minetest-dev
21:39 proller joined #minetest-dev
21:49 paramat left #minetest-dev
21:49 Megaf Thats fantastic! https://github.com/minetes​t/minetest/pull/1732/files
21:57 mos_basik joined #minetest-dev
21:58 mos_basik joined #minetest-dev
22:28 mos_basik joined #minetest-dev
22:29 shadowzone joined #minetest-dev
22:29 shadowzone joined #minetest-dev
22:31 blaise joined #minetest-dev
22:41 shadowzone left #minetest-dev
23:06 blaise joined #minetest-dev
23:24 paramat joined #minetest-dev
23:42 alexxsss joined #minetest-dev
23:46 kaeza is anyone interested in adding crouching to minetest?
23:51 kaeza read: I'm working on it, but no point in coding it "properly" if the answer is "no"
23:58 paramat i can't see a need for it

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