Minetest logo

IRC log for #minetest-dev, 2015-06-28

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

All times shown according to UTC.

Time Nick Message
00:10 proller joined #minetest-dev
00:39 Puma_rc joined #minetest-dev
01:11 Wuzzy joined #minetest-dev
01:12 blaise joined #minetest-dev
01:44 RealBadAngel joined #minetest-dev
02:02 RealBadAngel hmmmm, here?
02:02 hmmmm yes...?
02:05 RealBadAngel we have found reason for this "massive regression"
02:06 * VanessaE grabs the popcorn...
02:06 RealBadAngel kinda funny. est31 was using debug build...
02:20 RealBadAngel anyway, going to push #2848, any objections?
02:20 ShadowBot https://github.com/minetest/minetest/issues/2848 -- Minimap update by RealBadAngel
03:01 FR^3 joined #minetest-dev
03:07 FR^3 joined #minetest-dev
03:17 FR^3 joined #minetest-dev
03:59 FR^4 joined #minetest-dev
04:08 FR^4 joined #minetest-dev
04:18 FR^4 joined #minetest-dev
04:25 hmmmm RealBadAngel:  having a thread that does nothing too important and yet using 100% cpu even in a debug build is UNACCEPTABLE
11:30 loggingbot_ joined #minetest-dev
11:30 Topic for #minetest-dev is now Minetest core development and maintenance. 0.4.13 release scheduled for August 10 2015. Chit-chat goes to #minetest. Consider this instead of /msg celeron55. http://irc.minetest.ru/minetest-dev/ http://dev.minetest.net/
12:11 red1 joined #minetest-dev
12:43 SopaXT joined #minetest-dev
12:47 Wuzzy paramat, in GitHub you said something about underground biomes. do you want to explain?
13:07 prozacgod joined #minetest-dev
13:45 H-H-H joined #minetest-dev
13:56 Elinvention joined #minetest-dev
14:41 rubenwardy joined #minetest-dev
14:42 Taoki RealBadAngel: Compiling now to try out your minimaps. Thank you for this great feature :)
14:43 kilbith finally an interesting feature has come out :)
14:43 Taoki Indeed. Waiting for more noticeable features to happen.
14:44 Taoki The GIT is thankfully active. But it's usually changes that add nothing new, or don't improve performance noticeably.
14:46 kilbith yeah, crytography for protecting your apples in your chest is not that interesting from a player's standpoint
14:46 Taoki Interestingly enough, however, it doesn't work for me at all
14:46 kilbith fortunally we have RBA here for make this game worth of the name
14:47 Taoki Even with enable_minimap = true in minetest.conf
14:47 Taoki Yep :)
14:47 kilbith MT would fall into the boringness without him, sadly
14:48 rubenwardy Taoki, f9?
14:48 Taoki Aha! Thanks
14:48 Taoki Can you make it turn on by default automatically?
14:48 rubenwardy I don't know
14:50 Taoki If one is added, my game might enable this by default
14:51 Taoki One question however: How does it deal with being in deep underground biomes, or high up on floating islands... for biomes that do either?
14:52 Taoki Also, how can it be re-positioned. I would really like a way to move it in another corner on the screen... this might interfere with some HUD's
14:53 RealBadAngel Taoki, minimap setup via lua will be aviable in the future
14:53 RealBadAngel mods will be able to control it
14:53 Taoki RealBadAngel: Awesome. That's what I was thinking about, yeah
14:53 kilbith maybe craftable and placable in the future, also
14:53 kilbith as a separate node
14:53 RealBadAngel yes, thats what im thinkin about
14:54 Taoki RealBadAngel: More specifically: Only have the code generate the texture. Have mods be able to use it. On the HUD, or even on in-world paintings!
14:54 Taoki Yep ^^
14:54 RealBadAngel that would be nice, big billboard around the spawn with world map
14:54 Taoki Yep
14:55 kilbith but there are higher priorities atm
14:55 kilbith lighting, vbo, etc.
14:55 Taoki RealBadAngel: Also, I know you probably have a lot on your hands, but since you're most experienced with video stuff: Can you please, please, please try to be the hero that fixes this year old issue? https://forum.minetest.net/v​iewtopic.php?f=5&t=6418 (yes, I'm talking about the VBO code)
14:55 Taoki Woah... kilbith just mentioned VBO before I typed that o_o
14:55 RealBadAngel im playing with VBO right now
14:56 Taoki Of all coincidences xD
14:56 kilbith it being worked on
14:56 Taoki Woah, nice!
14:56 Taoki Again, of all coincidences. Can I read people's minds? ;)
14:56 RealBadAngel good is that it is working
14:56 RealBadAngel bad is that its gonna work flawlessly only when shaders enabled
14:56 Taoki RealBadAngel: But thank you then! I really want the FPS improvements it brings, they are very important. If only it wasn't for that cursed memory usage...
14:56 kilbith particles and lighting should be managed natively by irrlicht also
14:57 Taoki Eh... it can be shaders-only, don't mind. Anything as long as it works :)
14:57 Taoki kilbith: Agreed. Hardware is lighting is planned IIRC
14:57 Calinou we can easily wait 2 years for hardware lighting
14:57 Taoki I think VBO is a priority however. FPS and performance are a priority
14:57 Taoki Calinou: Sure.
14:58 Taoki Not much more than 2 years though :)
14:58 Taoki Anyway BBL
15:34 est31 joined #minetest-dev
15:37 RealBadAngel hi est31
15:37 RealBadAngel with those semaphores you forget bout one thing
15:37 RealBadAngel you let minimap update only when new blocks were loaded
15:39 est31 your fix is not needed, I think
15:40 est31 have you tested it?
15:40 RealBadAngel ofc its needed
15:40 RealBadAngel without it map will never update
15:40 RealBadAngel i mean without new blocks
15:40 est31 yes
15:41 RealBadAngel and thats wrong
15:41 est31 but usually a map update is a "new block"
15:41 RealBadAngel no
15:41 RealBadAngel map is being made out of cached blocks
15:42 RealBadAngel you havent get the idea how it works
15:42 RealBadAngel cached blocks are not the map
15:43 RealBadAngel theyre just data for mapper, and you have disabled map generation just
15:48 RealBadAngel anyway, you dont need now fancy gimmicks with texture
15:49 RealBadAngel i couldnt understand why you would ever need that, until i realized that you have broken map generation ;)
15:51 selat joined #minetest-dev
15:52 hmmmm joined #minetest-dev
15:55 rubenwardy Hi, I'd like reviewers for #2094 and/or #1748
15:55 ShadowBot https://github.com/minetest/minetest/issues/2094 -- Fix offset being ignored by inventory bar HUD by rubenwardy
15:55 ShadowBot https://github.com/minetest/minetest/issues/1748 -- Add lua errors to error dialog by rubenwardy
16:06 est31 ok RealBadAngel I agree the fix is needed
16:07 SopaXT joined #minetest-dev
16:18 est31 rubenwardy, 2094 is still waiting for sapier, isnt it
16:19 rubenwardy Sapier is the one that wrote the hud scaling code and knows how it works
16:19 rubenwardy but he's been AWOL for a while
16:20 rubenwardy What's it called when you copy a piece of code from one section to another, without understanding how it works?
16:20 est31 copy-paste??
16:20 est31 dunno
16:21 est31 either way I wonder how other people see this +1 rule for PRs from core devs
16:21 est31 for me, the author being a coredev counts as first +1
16:21 est31 hmmmm said its not how it was meant
16:22 rubenwardy It needs to be reviewed separately by other people, as the author may miss things
16:22 rubenwardy the whole clean pair of eyes thing
16:24 est31 I know, but the author of the PR is trusted, being a coredev
16:24 est31 I mean, then this maintainer thingy where you can push without any review is even worse
16:25 rubenwardy true
16:26 hmmmm okay
16:26 hmmmm so code peer reviews are usually carried out by two peers
16:26 hmmmm i.e. not the author himself
16:26 blaze joined #minetest-dev
16:26 hmmmm core developers are given more trust, they have commit privileges ffs
16:27 hmmmm but i don't see why there should be more complacency in reviewing their code
16:27 hmmmm core developers sure aren't falliable
16:27 hmmmm the trivial exception should be obvious
16:27 hmmmm the maintainer exception doesn't mean without review
16:27 hmmmm you should at least announce what's going on in the channel
16:28 hmmmm completely new features should probably have more review, like I would say 3 people
16:28 est31 hmmmm, all nice, but we have 95 open pull requests
16:29 est31 I still miss a second +1 from the game maintainers
16:29 est31 you cant say "I want 1000 people to review this" if none of them are active
16:29 hmmmm i know
16:30 hmmmm there's simply not enough active people
16:30 est31 if we had more, I'd agree to your ideas, but now we can't afford them
16:30 hmmmm so which is it?  compromise quality?
16:31 hmmmm i tend to value quality more than snazzy new features
16:31 hmmmm a lot of developers got upset with this choice and this is how freeminer was born
16:31 hmmmm less oversight, less review, less rigor, and as a result, less stable
16:31 est31 quality is important too, yes.
16:32 hmmmm a lot of people are already upset with how slow and unstable minetest is
16:32 hmmmm we don't need to make the problem worse
16:32 est31 a lot of people are already upset with how slow development is
16:32 hmmmm there's like less than one person working on minetest at any given time
16:33 est31 I try hard to reduce the number of open prs
16:33 hmmmm there are a lot of people who like to play video games but not a lot of people who are able to code them
16:33 hmmmm est31, let's go through some pull requests
16:33 hmmmm are you free for the next hour or so?
16:33 est31 yes
16:34 est31 I see it this way, if there is a problem with a merged PR, then those core devs who +1ed it (and the author) are responsible to fix the issue
16:34 hmmmm do you think that encourages people to +1
16:34 hmmmm or hide
16:34 est31 and btw, if we couldn't have fixed the 100% cpu problem yesterday, a git revert would have been justified
16:35 hmmmm I still think there's room for optimization there
16:35 hmmmm and I mean no offense to RBA, but knowing him, he's never going to perfect the feature
16:35 est31 thats what we have master for
16:36 est31 its not perfect yes, but already in a playable, acceptable state
16:36 johnnyjoy joined #minetest-dev
16:36 hmmmm if things don't get done the right way the first time around, nobody goes back to fix them
16:36 est31 also, which problems do you see with minimap?
16:37 est31 I think if we see master as a holy temple where only 100% perfect code can be commited, master is dead.
16:38 est31 I dont want to blindly add any pr proposed
16:38 est31 like freeminer
16:38 hmmmm well I don't know if it's a problem for sure but it seems to me like the consumer could get oversaturated
16:38 est31 eg they added gsmapper
16:38 est31 horribly laggy
16:38 hmmmm so
16:38 est31 we didnt
16:39 hmmmm we want developers who add things to follow through with their creations and make sure it's 100% perfect
16:39 est31 and rba rewrote a new minimap, and its much greater now
16:39 hmmmm but this isn't happening because polish is booooring
16:39 hmmmm what's fun is making new buggy code with awesome cool new features
16:40 hmmmm if this were the corporate world people would need to make their own code 100% perfect because their food depends on it
16:40 hmmmm we don't have that kind of coersiveness
16:41 est31 it depends on where you are in corporate world
16:41 hmmmm can you think of some way to encourage developers to develop complete solutions
16:41 est31 I agree we shouldn't rush things
16:42 est31 but very few people can write acceptable c++ code, and even fewer can polish it
16:42 hmmmm what I would like to know is how we're doing relative to other open source projects of the same type, with the same amount of users, and the same amount of developers
16:42 est31 I don't think we should say "sorry you cant contribute" for those who only write acceptable code
16:43 hmmmm because these are problems inherent with the development model
16:43 hmmmm maybe minetest is actually doing fantastic
16:43 hmmmm the Linux kernel is something that's quite universal
16:43 hmmmm it's useful to everybody
16:44 est31 but certainly not our size
16:44 hmmmm it has so many developers that they can afford to say "sorry you can't contribute" unless their code is stupendous
16:44 hmmmm so, is this some kind of open source project darwinism at work?
16:45 hmmmm the most inherently useful/universal projects get the highest quality development and the most attention?
16:46 est31 still it doesn't mean our code can't be polished
16:46 est31 SN likes to polishh things
16:46 hmmmm only the people who /care/ are the ones going to polish
16:46 hmmmm nothing is likely to come of this talk
16:47 hmmmm the problems are obvious, the solutions are not
16:47 rubenwardy How can we motivate people to do bug fixing and polishes etc?
16:48 hmmmm by its very nature, minetest attracts complete novices to coding
16:48 est31 I think, as long as we have acceptable master, where there are no side effects of new features, everything is fine
16:48 hmmmm the more advanced the developer, the less likely a bug fix or polish will be necessary to begin with
16:48 est31 polishing can be done later, and be done by other people who add the features
16:48 hmmmm who are the 'other people'?
16:48 rubenwardy I mean generally, as in bug fixing existing code.
16:48 est31 those who care, hmmmm.
16:48 est31 as you said
16:49 hmmmm there aren't enough of those types
16:49 rubenwardy joined #minetest-dev
16:49 est31 unfortunately
16:49 hmmmm there's me, you, zeno, SN,
16:49 hmmmm and that's about it
16:49 est31 yea :(
16:49 hmmmm I took the approach of being the polish-er for some guy's new feature
16:49 hmmmm that's the result of Hud
16:50 hmmmm err, that's how Hud came about
16:50 kilbith not only the core-devs does the dirty work fyi
16:50 rubenwardy I'd love to be able to help more, I just don't have the experience. :(
16:50 hmmmm Hud was actually nicely polished for a while before people started tacking more stuff onto it
16:50 hmmmm but it took the effort of two people at the same time
16:51 hmmmm and no doubt it's more frustrating for the other person
16:51 est31 i think this is a key
16:51 est31 everybody should have fun
16:51 est31 this is a free time project, no corporate project
16:52 est31 and even there its contraproductive to frustrate people
16:52 hmmmm i used to tell people that open source code is of the same or higher quality than proprietary code
16:52 Calinou it is still true :-)
16:52 hmmmm that's not true in the general case i think
16:52 hmmmm because look at the motivations at play
16:53 Calinou writing bad code at work is rarely frowned upon
16:53 hmmmm it might be true for linux or whatever
16:53 jin_xi joined #minetest-dev
16:53 Calinou maybe in some startups, but that's the exception, not the norm
16:53 est31 hmmmm, not every corporation makes code as polished as yours
16:53 est31 at least I think
16:53 hmmmm and not every corporation makes code as polished as my corporation does
16:53 hmmmm :\
16:54 hmmmm bleh
16:54 hmmmm i think it's impossible to know for sure
16:54 rubenwardy Not every open source project makes code as shitty as mine
16:54 hmmmm everybody is able to view all open source projects
16:54 hmmmm but only a very select subset of people are able to see a very select subset of corporate code
16:55 blaze joined #minetest-dev
16:56 est31 yea
16:56 hmmmm alright, starting right now, I'd like as many people as possible to participate in the PR clearing
16:56 est31 ok
16:56 hmmmm starting with the oldest PR first: #816
16:56 ShadowBot https://github.com/minetest/minetest/issues/816 -- Add wield light by Zeg9
16:56 est31 next hour is free for me
16:57 hmmmm so this PR did a great job at what it's supposed to do
16:57 hmmmm the primary hangup was that people were calling it unfair
16:58 est31 lemme see
16:59 est31 yes, its one of the design descisions that shaders may not affect gameplay
17:00 est31 what we need is a new "wield light" "layer"
17:00 est31 example of low quality vs feature
17:02 est31 ok what about the next open one #943
17:02 ShadowBot https://github.com/minetest/minetest/issues/943 -- Increase the amount of particles that breaking a node generates by ShadowNinja
17:02 hmmmm wait
17:02 est31 this one depends on irrlicht particles
17:02 hmmmm what was the resolution for the last one
17:03 est31 "needs improvement"
17:03 est31 what do you think
17:03 hmmmm so you think it should be added
17:03 kilbith est31, particles are effectively slowing down the game when there are in high number
17:03 est31 no
17:03 hmmmm the feature itself
17:03 est31 yes its good as feature
17:03 hmmmm when that PR first came around I supported it
17:04 hmmmm it's just that there were more naysayers than supporters
17:04 hmmmm so it sat
17:04 hmmmm the primary hangup is that it's unfair, but fairness is not determined by developers of the core, but rather the mod makers
17:04 hmmmm the subgame
17:04 est31 you didnt get a supporter (of the current state) with me :)
17:04 est31 unfair?
17:05 hmmmm unfair because you don't have to create torches to light up caves
17:05 hmmmm just one
17:05 est31 ah that
17:05 hmmmm I think that piece of code should be kept for future reference and historical reasons but
17:05 hmmmm wield lighting and things like that will be part of client side scripting if our security polices allow
17:05 est31 here I agree with you, its not a matter of core to decide
17:05 hmmmm brb
17:06 est31 but subgame
17:06 est31 server should send to client whether it has wield lightning
17:06 est31 either with client side shaders, or a simple flag in protocol, or whatever
17:07 est31 while I myself am for the flag
17:08 TeTpaAka joined #minetest-dev
17:09 hmmmm err...
17:09 hmmmm hold on, i'll be back in like 15 minutes
17:09 est31 ok
17:20 Taoki RealBadAngel: An idea has been on my mind recently. Since you are the shader developer, it's probably you I should bring it up to and ask about...
17:21 Taoki RealBadAngel: I was wondering whether Minetest could make use to one of those 2D filters used to smooth out pixelated textures or games. You see them on GameBoy Advance or DOS emulators. Basically, they take a large pixelated image and rounden all of the pictures. Such filters have names like SAI, or Super Eagle.
17:21 Taoki http://i.imgur.com/eo3IJ.png
17:21 Taoki http://gerbilsoft.soniccenter.org/gens/ma​nual/images/render/render-super-eagle.png
17:21 Calinou yeah, HQX or such
17:21 Taoki Some examples. The games in those screenshots are normally pixelated, like Minetest textures. Yet the filter makes them smooth like that
17:21 Calinou those are pretty expensive, and it doesn't bring much to add them
17:22 Taoki I see
17:22 Taoki I was thinking that with x16 textures, such a filter could be used to soften them and make them look more artistic
17:23 Taoki http://filthypants.blogspot.ro/2012/03​/xbr-vs-hqx-interpolation-filter.html This descripes the idea in more depth
17:26 Taoki https://en.wikipedia.org/wiki/Hqx Even better, to get an idea of what and how
17:27 Elinvention what's the problem with hardware lightning in minetest?
17:27 Taoki Elinvention: No one's gotten around to implementing it yet.
17:28 Elinvention Minetest is based on Irrlicht right? Irrlicht doesn't support it?
17:29 est31 first: yes, second: dunno
17:29 Calinou hardware lighting is also quite slow
17:29 Calinou even for today's hardware
17:29 Calinou (that is, realtime lighting)
17:30 Taoki Elinvention: It does. We just aren't using it yet.
17:30 Taoki Also seems someone already wrote hqx glsl filters. https://github.com/Armada651​/hqx-shader/tree/master/glsl
17:31 Taoki https://www.shadertoy.com/view/MslGRS Cool, another example :o
17:36 Elinvention If hardware lightning were used, the shaders wouldn't have to be on to see the light, isn't it?
17:37 Elinvention sorry lighting
17:44 Taoki Yes, that's not shader related
17:51 Garth joined #minetest-dev
18:10 OldCoder joined #minetest-dev
18:14 CraigyDavi joined #minetest-dev
18:21 H-H-H joined #minetest-dev
18:38 TeTpaAka joined #minetest-dev
19:06 prozacgod joined #minetest-dev
19:24 AnotherBrick joined #minetest-dev
19:24 est31 hmmmm, back?
19:24 hmmmm yeah
19:25 est31 ok, so the oldest pr needs enhancement, and some call it unfair
19:25 hmmmm alright so my recommendation for 816 is that the code should be kept for reference purposes but it won't be merged in its current form at all
19:25 est31 yes
19:25 hmmmm the reason why it's unfair is because fairness is defined by the subgame, not the developers
19:25 hmmmm so it's a client-side modding thing
19:26 est31 or at least it should rely on some value sent by the server to the client
19:26 est31 and not settable by the client.
19:26 hmmmm absolutely
19:26 est31 ok #943
19:26 ShadowBot https://github.com/minetest/minetest/issues/943 -- Increase the amount of particles that breaking a node generates by ShadowNinja
19:26 hmmmm if we do decide that running shaders downloaded from a server is safe enough, it'll be implemented that way
19:26 hmmmm so I think we should close it
19:26 hmmmm btw
19:27 hmmmm we need to end each of our discussions with a final decision to keep open and add a comment to the PR, or close it out
19:27 est31 doesn't "keep around for reference purposes" mean keeping open?
19:27 hmmmm keep it somewhere
19:27 hmmmm make a branch perhaps
19:27 hmmmm i don't know, but it's not going to be merged so it should be closed
19:27 hmmmm #943
19:27 ShadowBot https://github.com/minetest/minetest/issues/943 -- Increase the amount of particles that breaking a node generates by ShadowNinja
19:28 RealBadAngel hmmmm, i rather take care of my code, when given feedback im able to fix things
19:28 hmmmm what do you think
19:29 RealBadAngel see what happened just a few minutes ago in main
19:29 hmmmm the particle amount should be a setting
19:29 TeTpaAka joined #minetest-dev
19:29 hmmmm not some hard coded thing
19:29 kilbith particles in high amount is FPS killer, VanessaE please told them
19:29 kilbith (example : shower head's particles)
19:29 hmmmm that's because the current particle implementation sucks
19:29 kilbith exactly
19:29 hmmmm so...
19:29 hmmmm want to close it out?
19:29 est31 yes
19:29 hmmmm alright
19:30 hmmmm close #943
19:30 ShadowBot https://github.com/minetest/minetest/issues/943 -- Increase the amount of particles that breaking a node generates by ShadowNinja
19:30 est31 its too small for it
19:30 hmmmm make sure you actually add a reason why you're closing it too btw
19:30 hmmmm or do you want me to
19:30 RealBadAngel hold on a second with those particles
19:30 est31 close it, I am doing it for 816
19:30 RealBadAngel we are talking about a few particles when diggin a node
19:31 RealBadAngel its absolutely not a huge ammount and it doesnt affect fps
19:31 RealBadAngel its not moddable stuff
19:31 RealBadAngel i did the same with digging tweaks
19:32 kilbith RBA, try out the homedecor shower head's particle with 60 particles/s
19:32 RealBadAngel cracks are far worse for the engine (by triggering several times mesh update) than those few particles
19:32 kilbith you'll see the bad effect
19:33 RealBadAngel damn, we are not talking about shower now
19:33 kilbith it's the same particles
19:33 RealBadAngel just a few more particles when digging
19:33 kilbith and it's already really BAD with 60 particles/s
19:33 RealBadAngel cracks are worse
19:33 kilbith FPS drops as -15 FPS
19:34 RealBadAngel with what?
19:34 kilbith a old GPU
19:34 RealBadAngel what is causing that drop i ask
19:34 hmmmm a crack animation triggers a mesh update?
19:34 kilbith when you turns on the shower head flowing 60 particles/s
19:35 RealBadAngel hmm suprised?
19:35 hmmmm it shouldn't do that
19:35 RealBadAngel kilbith, that issue wasnt about the shower
19:35 kilbith but i talk about the particles not the shower itself
19:35 kilbith VanessaE ^
19:36 RealBadAngel kilbith, dont mix issues plz
19:36 RealBadAngel hmmmm, i said about that many times and tried to replace those damn cracks with particles
19:36 RealBadAngel but cracks seems to be kinda freaky religion here
19:36 est31 #1685 is too ugly
19:37 ShadowBot https://github.com/minetest/minetest/issues/1685 -- Digging animations tweaks. by RealBadAngel
19:37 RealBadAngel ugly?
19:37 RealBadAngel oh thx :P
19:37 est31 jugding from the video, i dont like it https://www.youtube.com/watch?v=wjG1bUBFPB0
19:37 est31 lol
19:37 RealBadAngel fuck the vid
19:38 RealBadAngel try the feature before trashing it
19:38 hmmmm yeah I don't like it either...
19:38 RealBadAngel so live with cracks :P :)
19:39 hmmmm there should be a technically better way to replace a single texture in a mesh with a different variant
19:39 RealBadAngel theyre even more ugly, doesnt fit all drawtypes, eat memory and lag engine
19:39 RealBadAngel long live the cracks!
19:39 est31 heh
19:40 hmmmm don't change a part of minetest that everybody likes because you personally can't figure out a way to make them more efficient
19:40 RealBadAngel nobody even noticed that this feature has an option to turn it on and off
19:40 Krock http://i62.tinypic.com/2qkt6j5.jpg I want cracks too
19:40 RealBadAngel so one can have cracks, another highlighting plus particles
19:41 hmmmm I think if we are hoping to generalize minetest more into an engine than a minecraft clone with some lua scripting, it's essential to somehow abstract the crack animation away from digs
19:41 kilbith RBA, the trigger on particles is the same whether you're digging a node or activate a particlespawner
19:41 kilbith so it's the same issue
19:41 RealBadAngel you dont dig 100 nodes at the time
19:42 RealBadAngel so it cannot cause an issue like lets say 100 showers :P
19:42 MinetestForFun joined #minetest-dev
19:42 kilbith how many time per second you punch a node when holding left click ?
19:43 kilbith it's lotsa of particles
19:43 RealBadAngel if that digging effect is causing fps drop, im chinese ballet dancer :P
19:44 RealBadAngel and hmm, est31 how easy for you to drop the feature without even trying how it works and feels
19:44 RealBadAngel thats common
19:44 Krock it's surely causing a minimalistic fps drop :P
19:44 hmmmm alright
19:44 RealBadAngel "screenshot is enough"
19:45 RealBadAngel bullshit
19:45 hmmmm so I just closed #943
19:45 ShadowBot https://github.com/minetest/minetest/issues/943 -- Increase the amount of particles that breaking a node generates by ShadowNinja
19:45 hmmmm onto #958
19:45 ShadowBot https://github.com/minetest/minetest/issues/958 -- Add player list, viewable by holding Tab by sfan5
19:45 hmmmm ahh, this one... I remember this one
19:46 hmmmm this was implemented back when we couldn't change fonts or font colors and so the text was white on light gray and was practically impossible to read
19:47 hmmmm it could be fixed and merged, but do we really want this to be a part of the core?  this is the kind of thing that could be trivially implemented in a mod
19:47 hmmmm what do you guys think?
19:49 est31 yea a mod should do this
19:49 TeTpaAka This should be made by a mod, imho. The only problem for a mod would be the key binding.
19:49 TeTpaAka But this could be fixed by client side modding.
19:51 hmmmm client side modding has been the answer for a couple of these mods now
19:51 hmmmm I'm going to boost the priority of client side modding
19:51 hmmmm this is going to be what I work on after fixing some immediate bugs
19:51 hmmmm we'll get the keybinding and screen drawing/hud features out first
19:52 hmmmm formspec is going to be more difficult to replace so that'll come later rather than sooner
19:52 est31 ok
19:53 hmmmm but yeah I don't agree with adding things like this to the core
19:53 hmmmm it's a core, not a game
19:53 hmmmm the game is the collection of mods
19:56 * VanessaE wanders in
19:57 est31 what bugs?
19:58 hmmmm I'm going to finally fix the damage flashing bug
19:58 VanessaE kilbith is right about the particles though - whether they are spawned by a dig action, single-particle actions (as in the "snowdrift" mod), or as spawners like homedecor's faucets and showerhead, the more particles there are, the slower the FPS.  Too late now since you already did it, but I sadly agree to closing SN's PR.
19:58 hmmmm there's also the player texture bug
19:58 hmmmm basically, nerzhul leftovers that he never bothered with
19:58 VanessaE hmmmm: you talking about the fake damage in creative mode?
19:58 kilbith ^ RBA read that
19:58 hmmmm yes
19:59 hmmmm i know the reason for that
19:59 est31 player texture?
19:59 VanessaE est31: sometimes the player texture comes out flat
19:59 jin_xi lots of particles with collision detection. its expensive
19:59 hmmmm tl;dr the server sends a DIEPLAYER event even if damage is disabled
19:59 VanessaE instead of wrapped around a model, or comes out as "the green guy" for no reason
20:00 hmmmm so what do you want to do
20:00 hmmmm close out 958?
20:01 est31 yes
20:01 hmmmm we should be keeping track of which features are getting re-written using client side modding
20:01 est31 there is http://dev.minetest.net/Client_scripting_plans
20:01 hmmmm there it is
20:02 hmmmm i was literally about to ask you for the link
20:02 est31 heh
20:02 est31 It has mostly stuff about the general construction
20:02 est31 not about the feature
20:02 est31 s
20:02 hmmmm I am adding a new section
20:02 est31 but certainly a good place to put them
20:02 est31 ok
20:05 hmmmm maybe we should add an issue to make the number of particles settable per-nodedef
20:06 hmmmm so digging a sand node would make lots of particles whereas others maybe not so much
20:06 hmmmm an issue so we remember about it at least
20:06 est31 good idea
20:07 prozacgod joined #minetest-dev
20:07 VanessaE offtopic, sorta:  https://forum.minetest.net/viewtopic.ph​p?f=11&t=3933&p=183022#p182987
20:08 VanessaE (summary:  it would be nice if a server could detect settings on the client, for client-side configuration of a server-side mod)
20:09 VanessaE hmmmm: as I recall, sfan5's old particles mod could (if extended) do exactly that
20:09 hmmmm hmm
20:09 VanessaE but it used entities to spawn particles, so it was even slower, and laggy as well
20:10 hmmmm you have me wondering now
20:10 VanessaE ?
20:10 hmmmm see, I had the idea of maybe using client side scripting for all special visual effects
20:10 hmmmm like remember the light aura around light sources i implemented
20:10 hmmmm instead of making that an attribute of the node
20:11 hmmmm that should be a scripted special effect to draw each time that type of node is rendered
20:11 hmmmm except in some manner that makes it fast
20:11 VanessaE I remember that light aura
20:11 hmmmm in the same way, could a local-only particle spawner be created on_dig?
20:11 VanessaE in fact I was just thinking about that last night, wondering if it could be implemented again
20:11 hmmmm i wouldn't hard code the effect in
20:12 est31 good idea
20:12 hmmmm it would have to be made generic somehow
20:12 hmmmm and keep in mind this also needs to be fast
20:12 hmmmm i.e. register things, not execute lua on a callback
20:12 hmmmm and it needs to not suck
20:13 VanessaE well this is one of those times when it would be pretty obvious how the API should work at least
20:13 VanessaE register_node(foo, { client_on_dig = function(blah) ; do particles ; end })
20:14 VanessaE i.e. like a callback, but said code would be uploaded to the client at init, to be executed later
20:15 hmmmm that's how it's already going to work
20:15 VanessaE well at least we agree on THAT much :P
20:16 hmmmm I don't know, maybe digging a node isn't hot enough of an event to have a registration for
20:16 hmmmm but node rendering effect should be registered
20:16 VanessaE why not?
20:16 hmmmm because
20:16 hmmmm you hardly ever dig a node
20:16 VanessaE we have callbacks for 1927423 other things
20:16 hmmmm so there will be two callbacks
20:16 hmmmm start_dig
20:16 VanessaE oh no, believe me that's not true :)_
20:16 hmmmm end_dig
20:16 hmmmm it'll be like mouse_up and mouse_down
20:16 VanessaE that's not too bad an idea
20:16 hmmmm end_dig will have some kind of boolean if the dig completed
20:17 VanessaE good idea
20:17 hmmmm well
20:17 hmmmm i'm just spouting obvious ideas
20:17 VanessaE but I mean, we have on-construct, on dig, after dig, and a few others
20:17 sfan5 >on dig, after dig
20:17 est31 gtg
20:17 sfan5 wouldn't on dig be equivalent to on punch then?
20:17 VanessaE I meant on punch rather
20:18 hmmmm oh
20:18 hmmmm sfan5:  perhaps so
20:18 hmmmm except on_punch would be much more generalized than on_dig
20:19 hmmmm est left
20:19 hmmmm would the rest of you like to continue looking at old PRs?
20:20 kilbith est31 ^
20:20 hmmmm he just said he left
20:21 VanessaE old PRs is a good idea.
20:21 hmmmm the next one is kind of a doozey
20:21 hmmmm i'd like to skip it for now...
20:21 VanessaE ?
20:26 est31 back?
20:26 est31 back!
20:26 VanessaE BACK!
20:26 VanessaE :)
20:26 VanessaE hmmmm: which one are you talking about?
20:26 hmmmm meta set nodedef
20:26 VanessaE eek
20:26 hmmmm i would like to merge that eventually
20:27 VanessaE that's been wanted for like, forever, but one thing that occurred to me recently is, what's the worst-case scenario for a map with a fuckdton of nodes so-modified?
20:27 VanessaE I mean, double the size?  triple?
20:27 hmmmm i'm not really sure
20:27 hmmmm it's going to be much larger, definitely.
20:27 est31 it depends on how the format is
20:27 hmmmm i thought you left est
20:27 hmmmm :)
20:27 est31 yea
20:27 est31 <est31> back!
20:28 hmmmm oh
20:28 hmmmm rrrr
20:28 hmmmm anyway skipping #1118 for now... like I said, it's a doozey.  but I want it at least, and it needs a rebase
20:28 ShadowBot https://github.com/minetest/minetest/issues/1118 -- Meta set nodedef by Novatux
20:29 hmmmm #1188 looks pretty good apart from some minor style issues
20:29 ShadowBot https://github.com/minetest/minetest/issues/1188 -- getTime refactoring by Selat
20:29 hmmmm how do you guys feel about separating porting.cpp things into separate files based on 'topic'?
20:29 est31 and that it needs a rebase
20:29 est31 if we do so, we should add a new directory
20:29 est31 only for porting
20:30 est31 but ok
20:30 hmmmm yeah
20:30 hmmmm so this is going to add get_time.cpp or something
20:31 hmmmm porting/get_time.cpp?
20:31 hmmmm instead of "porting" why not something more immediately clear such as "platform-dependent code"
20:31 hmmmm that should be in a separate issue, though.
20:32 est31 yes
20:32 est31 it should still be inside the porting namespace
20:33 hmmmm what arguments are there for having it namespaced?
20:33 est31 because its only a platform dependent helper code
20:33 est31 basically what porting is
20:37 Nihao Hello, protectors of the core cubes :)
20:37 Nihao I have a question about loss of material on exchange:   http://pastebin.com/awvFdh14
20:37 Nihao Probably the same as    https://github.com/minetest/minetest/issues/2028
20:37 Nihao It kind of "disturbs" (!) the gameplay and seems to appear with Barter Table and chests at least.
20:37 Nihao Is it possible to fix that?
20:38 est31 it is if you know how to reproduce
20:39 Nihao est31: Sorry, I do not reproduce it, but it crashes the server 10 tiles a day.
20:39 Nihao times*
20:39 Nihao Just crashed again....
20:40 est31 it is possible that a player used this to volunarily crash the server
20:40 est31 perhaps
20:40 Nihao How to know?
20:41 Nihao Can you instrument the code?
20:41 Nihao To have more data in the logs
20:41 est31 Nihao, just try a debug build
20:41 est31 and open it in gdb
20:42 Nihao With 25 people?
20:42 est31 at least its no nrz bug
20:43 Nihao :) seems to be good news. What does it mean?
20:43 est31 that its old
20:44 Nihao nrz
20:44 Nihao It is old and has neve been fixed.
20:44 kilbith nrz = core-dev
20:44 Nihao Maybe a survivor.
20:45 Nihao Or a really hard one.
20:46 Nihao Or both, anyway, how can I instrumnt the code on the "productive" server to have more useful data in the logs?
20:46 est31 debug build
20:47 est31 gdb
20:47 est31 gdb has no real performance penalty
20:47 est31 or no big
20:47 est31 unlike valgrind
20:47 est31 that one is a biggie
20:48 Nihao OK, thanks est31, I have to think that over and will hopefully be back with more precise indications soon.
20:55 zat joined #minetest-dev
21:25 kilbith so the PRs clearing out has stopped ?
21:26 est31 seems so
21:27 hmmmm nope
21:27 hmmmm let's continue on
21:27 hmmmm let's make our goal to get below 90
21:27 hmmmm so what do you guys think about #1277
21:27 ShadowBot https://github.com/minetest/minetest/issues/1277 -- Update to Lua 5.2 by ShadowNinja
21:28 kilbith better picking 5.3 now maybe ?
21:28 hmmmm yeah, i was thinking the same.
21:28 hmmmm does anybody know the story of 5.3 compat with luajit?
21:30 est31 its nonexistent I think
21:30 est31 last time I checked
21:30 est31 and yes, we can switch to any version luajit supports
21:30 hmmmm the whole selling point of luajit is that you can drop it in and it'll work
21:30 est31 but only fully supports
21:30 est31 and other way round too
21:30 hmmmm right
21:30 hmmmm it needs to be a 1:1 replacement
21:31 est31 yes
21:31 hmmmm what do you think we should do
21:31 hmmmm keep open or close?
21:32 est31 SN has invested quite alot of work
21:32 est31 it seems
21:32 est31 a large diff
21:32 est31 ah
21:32 hmmmm the majority of that is updating the builtin version of lua
21:32 est31 its only cped from lua source
21:32 hmmmm which is not builtin anymore
21:32 est31 yea
21:33 est31 really?
21:33 est31 what is in src/lua then
21:33 TeTpAaka joined #minetest-dev
21:34 hmmmm ah..
21:34 hmmmm maybe it doesn't use builtin by default then..?
21:34 hmmmm oh it was sqlite that got dumped
21:34 est31 what I hate about lua is that they break every single release
21:34 est31 yes
21:35 hmmmm why is there so much inconsistency between which libraries get bundled by default
21:35 est31 this is a thing a language shouldnt do
21:35 est31 ?
21:35 hmmmm sqlite isn't bundled, but lua is bundled
21:35 hmmmm why was this choice made
21:35 est31 dunno, before my time
21:36 hmmmm irrlicht isn't bundled
21:36 hmmmm so you can't make the argument that only essential libraries are bundled
21:36 est31 I dunno
21:37 hmmmm so hrmm
21:37 hmmmm updating lua to 5.2 - close out?
21:38 est31 http://lua-users.org/lists/​lua-l/2012-01/msg00787.html
21:38 est31 "o not hold out for a fully 5.2-compatible version of LuaJIT, if you
21:38 est31 are thinking of it. Mike Pall has made it clear that he dislikes and
21:38 est31 disagrees with many of the choices made by the Lua team for this
21:38 est31 version, and has no intention of moving away from a primarily
21:38 est31 5.1-based LuaJIT."
21:38 est31 cite that and close it
21:38 hmmmm luajit support is honestly more important than lua support
21:39 hmmmm in 99% of the cases people are going to use luajit
21:39 est31 yes
21:39 hmmmm (also that's an argument why I believe lua shouldn't be bundled by default any longer)
21:39 est31 but in the 1% where sb runs valgrind, things have to run smoothly too
21:39 hmmmm should we get rid of minigmp too, do you think?
21:39 est31 *shrug*
21:40 est31 if you get the windows build people to agree
21:40 hmmmm windows builds are a whole other level of pain
21:40 est31 OTOH, who is hurt by those files
21:41 hmmmm well for a while I tried taking over the windows build
21:41 hmmmm it was quite tough getting all the dependencies together
21:45 est31 so, the PR can be closed?
21:46 hmmmm yeah hold on
21:46 hmmmm i am closing it
21:46 est31 ok
21:50 est31 next one?
21:51 TeTpAaka2 joined #minetest-dev
21:53 hmmmm yeah
21:53 hmmmm #1220
21:53 ShadowBot https://github.com/minetest/minetest/issues/1220 -- Player information window by MirceaKitsune
21:54 hmmmm watch 5 more PRs get added tomorrow
21:54 est31 heh
21:54 hmmmm it's enough to make you say fuckit and give up
21:54 hmmmm hmm
21:54 hmmmm so this is formspec-based
21:55 hmmmm i like the idea behind it
21:55 est31 this one is better
21:55 hmmmm the implementation sorta sucks though
21:56 hmmmm the implementation is pretty good for what it is, a formspec, but the problem is that it uses formspec
21:56 hmmmm and you have to type a command to open it
21:56 hmmmm could we merge this together with the connected player list idea for client-side modding?
21:57 kilbith it's too specifical to be added in builtin imho
21:57 est31 if we go with our previous descision, we have to close this too
21:57 kilbith belongs to a mod
21:57 est31 ^
21:57 hmmmm maaaaaaaybe
21:57 est31 or minetest_game
21:57 hmmmm yeah
21:57 est31 I'd like to have it in minetest_game
21:58 hmmmm idea time
21:58 hmmmm what if basic player connection info gets displayed as part of a list in builtin client side modding
21:58 hmmmm and then other games can override the info screen
21:59 hmmmm so if you hold down tab on a game that doesn't implement this special info screen it'll just show basic information like name, connected time, latency
21:59 kilbith also no need a formspec here, /ping <player> would be sufficient
21:59 hmmmm but if you have special player information in your special game
21:59 hmmmm you can override the tab button to open /your/ player info display with the game-specific stuff
21:59 hmmmm yes? no?
22:00 est31 I dunno
22:00 est31 it gives this "default" feature
22:00 hmmmm the default thing could be part of minetest_game then
22:00 hmmmm well
22:00 hmmmm I don't know, we'll cross that bridge when we come to it
22:00 est31 do we have this at other places too?
22:00 hmmmm what now
22:01 est31 yea good
22:01 est31 it has issues of its own
22:01 est31 like minetest.setting_get("max_users")
22:01 est31 wrong info
22:01 hmmmm ohh :/
22:02 est31 (unrelated from whether to close or not)
22:02 est31 I'd suggest we close this too, and wait for client side scripting and a formspec replacement
22:02 hmmmm yeah
22:02 hmmmm we can do soo much better
22:03 hmmmm like don't get me wrong - i think this is great
22:03 hmmmm it's the best that it can be
22:04 hmmmm but the interfaces suck and it's like we're telling people to build a house with a saw and no hammer
22:04 est31 yea its dangerous to close PRs because "clientside scripting will make this all obsolete"
22:04 hmmmm right
22:05 hmmmm after this first batch that's what we'll get working on
22:05 est31 sfan5's solution is more proper
22:05 hmmmm this is pretty high priority now that i look at it
22:07 TeTpaAka joined #minetest-dev
22:07 hmmmm in fact gosh
22:07 hmmmm you know what
22:07 hmmmm a majority of these PRs are obvious client side scripting things
22:08 hmmmm I think there's a reason for this.  the Minetest UI is one particularily deficient area and it's incredibly visible, so people tend to focus on that
22:08 hmmmm i don't want to skip around but I think it's for the best
22:09 kahrl I have a different idea for the player list, maybe completely off base but it might work
22:10 kahrl essentially adding a tiny thing to the core that would allow doing the rest in a mod
22:10 kahrl the idea is to add a playerlist flag to functions like hud_add
22:10 kahrl hud items with that flag set will only be shown whenever tab is pressed
22:11 kahrl what do you think?
22:11 est31 I don't know how good the HUD api is
22:11 hmmmm it sucks
22:11 hmmmm just trust me, it sucks
22:12 hmmmm it works for now though, that's why people use it
22:12 kahrl maybe it's time to improve it :P
22:12 hmmmm I suppose your suggestion could work as a stopgap measure
22:12 est31 yes
22:13 hmmmm it's as simple as adding another attribute to hud items
22:20 kilbith #1328 can be closed as well.
22:20 ShadowBot https://github.com/minetest/minetest/issues/1328 -- Add 3D torches by BlockMen
22:20 kilbith the shape is hardcoded and it's saner to have a meshnode for maintainance / preferences.  belongs to subgames.
22:21 CraigyDavi joined #minetest-dev
22:22 est31 yes
22:23 est31 and this 2 pixel proble
22:23 est31 m
22:23 est31 or no
22:23 est31 I actually hate this "flat torch" thing
22:23 est31 thats hardcoded too
22:23 est31 this is at least 3d
22:23 est31 like a proper object
22:25 kilbith also, keep in mind that torchlike is not only used for torches
22:25 est31 but?
22:26 kilbith this code can fuck up nodes like lanterns or so
22:26 kilbith whatever that is not a torch
22:26 est31 they should use meshnodes
22:26 kilbith not necessarily
22:26 est31 its horrible to have 2d objects in a 3d world
22:27 kilbith same goes for other drawtype then
22:27 est31 the only exception being things like curtains
22:27 kilbith if you force the 3d here, it's no longer flexible
22:28 kilbith and easy to maintain
22:29 kilbith or alternatively this can be added as 'torchlike3d' drawtype
22:30 est31 the only torchlike node in homedecor is a cobweb
22:30 kilbith homedecor is not all
22:30 est31 which is perhaps even a good usecase
22:31 kilbith from a modder standpoint i'm strongly opposed to hardcode a 3d shape to torchlike
22:32 est31 it should never have been added
22:32 est31 same as for plantlike
22:33 kilbith but c55 liked the flatness of those drawtypes :)
22:33 kilbith it's just FOSS, you just have to draw pixels with gimp
22:33 kilbith s/FOSS/KISS
22:35 kilbith that is more flexible : https://github.com/minetest/minetest/pull/1871
22:35 est31 perhaps we should add a "flat" drawtype
22:35 est31 but very laggy I think
22:35 kilbith don't add a flat drawtype, the default torchlike should stick 2d
22:36 kilbith because then you would force people to update their mods
22:36 est31 its only a string value
22:36 est31 ok
22:36 kilbith really not, it belongs to a subgame as meshnode
22:36 est31 perhaps we should add a torchlike3d and a flat drawtype
22:36 kilbith no need to complicate things
22:37 est31 and then tell people to update via logging
22:37 est31 do that for one release, and then the next release remove it
22:37 kilbith for the history, it's me that requested to BlockMen to reopen this PR
22:38 kilbith i thought wrong
22:38 Taoki Ugh... a client Lua API is still being planned? Asking because it's one thing I am still not very sure about
22:38 est31 it is planned yes
22:38 Taoki Although, I guess, it is a good idea. Just worried it might make modding too complex and divided and weird
22:39 est31 flexibility always comes at this cost
22:39 est31 but running code clientside has huge advantages
22:39 est31 you just dont have the network lag
22:39 Taoki True, true
22:39 est31 and its distributed, and doesnt need server CPPU
22:39 est31 -P
22:40 Taoki Hope it can be implemented in a clean and non-weird way
22:41 kilbith kwoelkr's code is rather strong, no worries for that
22:42 zat joined #minetest-dev
22:46 cvtsx joined #minetest-dev
22:47 est31 ok lets skip that then
22:47 est31 hmmmm, next one?
22:48 hmmmm yep
22:49 hmmmm close #1328?
22:49 ShadowBot https://github.com/minetest/minetest/issues/1328 -- Add 3D torches by BlockMen
22:49 hmmmm this is something that clearly should be a meshnode but it's hardcoded
22:49 hmmmm i'd like 3d torches but it doesn't need to get added to the engine
22:50 est31 the torchlike drawtype has a bad name
22:50 kilbith yeah as i said, it's saner to use a meshnode for maintainance and fits the preferences
22:50 est31 doing "flat" is hardcoding too
22:50 hmmmm it is possible to wallmount meshnodes?
22:50 est31 I'm not sure
22:50 kilbith but flatness is KISS, you don't have to bother with Blender
22:53 est31 I'd merge the PR if 2 criteria are met: 1. make it torchlike3d, leave torchlike untouched (that one should be renamed to flat in the future, but thats other issue) 2. also work with higher density textures. just make the "middle" with 1/8 of the texture size
22:54 hmmmm yeah but it's hardcoded
22:54 est31 we will get more hardcoded drawtypes in the future
22:54 hmmmm that's so useless
22:54 est31 like cablelike
22:54 hmmmm ugh
22:54 hmmmm i would strongly prefer to shy away from hardcoded things
22:54 est31 otherwise you will have to register hundreds of connection combinations
22:55 hmmmm why can't this be done with a meshnode?
22:55 est31 as kilbith said, you can then make torches only with gimp, no blender needed
22:55 est31 we are a minecraft-like voxel game engine, no irrlicht
22:58 prozacgod joined #minetest-dev
22:58 hmmmm yeah but
22:58 hmmmm you can't do anything differently
22:58 kilbith hardcoded = no flexibility
22:59 kilbith against the master principle of MT
22:59 hmmmm you can make torches... only torches.  the only torches you'll be able to make are rectangular prisms jutting out of a wall at an angle
22:59 hmmmm the only flexibility you'd have is making the texture slightly different
22:59 est31 so, then remove torchlike altogether?
22:59 hmmmm and with such a small texture that'd mean the only thing you'd be able to do is change the color
23:00 est31 its the same as for torchlike.
23:00 hmmmm torchlike needs to stay around for reverse compatibility
23:00 hmmmm it wasn't the smartest idea, but it was the best they had back then
23:00 est31 we should deprecate it then
23:00 est31 with a warning
23:01 Sokomine joined #minetest-dev
23:01 kilbith est31, talking about MC, you realize that there are some flat drawtypes in MC ?
23:01 est31 which?
23:01 kilbith 'signlike'
23:02 kilbith hmm no, ladderlike
23:02 kilbith but they are named differently
23:02 est31 yea ladders are even more or less ok
23:03 kilbith ah, and the plants are all flats as well !
23:03 kilbith like our plantlike
23:03 kilbith so please listen to the modders for once
23:03 kilbith that's a pretty bad idea to force using 3d there
23:04 kilbith especially a determinated shape
23:04 est31 hmmmm, kilbith from what I find, we can't do right that torches blockmen proposes in that PR
23:04 est31 but only ones where there is no texture sticking out at the sides
23:05 est31 I think we should add this as a stopgap
23:05 hmmmm you realize it becomes permanent, right?
23:05 est31 until we have wallmounted meshnodes
23:06 hmmmm wait
23:06 hmmmm are you sure we don't have wallmounted meshnodes already?
23:06 hmmmm or anything that can replicate wallmounted?
23:06 hmmmm vanessae?
23:07 hmmmm est31: celeron seems to agree with me
23:08 kilbith http://irc.minetest.ru/min​etest/2015-01-03#i_4092174
23:08 est31 hmmmm, noted
23:09 est31 ok, so you can re-do it with meshnodes, but without flames sticking out at the sides?
23:09 hmmmm we should probably look into ways to make implementing 3d torches simpler using meshnodes
23:26 Sokomine regarding 3d torches: they may look nice but are usually horrible regarding gameplay. client prediction usually fails, so it feels like a very bad lag, and there are artifacts left when removing a torch. at least it was that way the last time i saw it used
23:26 Calinou meshnodes make client prediction possible, normally
23:26 Calinou using wallmounted parameter
23:26 Calinou nodeboxes too but nodeboxes can't be rotated (see carbone_torches)
23:27 est31 lets jump that pr
23:27 est31 nex one?
23:29 Sokomine Calinou: whatever. as long as it doesn't cause players to experience lag when placing a torch to get light it ought to be ok
23:46 Nihao left #minetest-dev
23:55 prozacgod joined #minetest-dev
23:59 Fritigern joined #minetest-dev

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