Minetest logo

IRC log for #minetest-dev, 2013-02-15

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

All times shown according to UTC.

Time Nick Message
00:02 ShadowNinja hmmmm: responsiveness seems better
00:14 hmmmm shadowninja, are you on windows?
00:19 celeron55 joined #minetest-dev
00:40 RealBadAngel hmmmm, http://pastebin.com/GCfzceWR
00:40 RealBadAngel cannot compile that branch
00:42 VanessaE hmmmm: check your last two commits, one of those is responsible
00:42 hmmmm ?
00:42 VanessaE I just tried to git pull + compile and got the same error as RealBadAngel
00:42 hmmmm weird....
00:44 VanessaE you bwoke it! :'(
00:44 hmmmm now why does this work for me
00:44 hmmmm anyway just put that label after END_DEBUG_EXEPTION_HANDLER(errorstream)
00:50 doserj joined #minetest-dev
00:50 RealBadAngel hmmm, btw i was reading the whole discussion regarding rotating nodes and how it should be done ideas
00:51 RealBadAngel answer is any other way is impossible from what i can see in code
00:52 RealBadAngel if cuboid would be rotated, it will end with rotating lights, shader effects and so on applied to all faces
00:57 VanessaE hmmmm:  just checked, commit 6981e3[...] seems to compile okay, so it's just that last one.
00:57 hmmmm i know exactly what it is...
01:06 hmmmm celeron, i do not understand the whole "add" and "avg" profiler thing, what do each of them do exactly?
01:06 hmmmm the code for it is.... quite... odd...
01:46 RealBadAngel hmmmm, after some testing i think it may be possible on the vertex level
01:47 RealBadAngel but not all of my code has to be dumped, that lookup table is necesary
01:47 hmmmm i haven't looked at it too much, but that's good
01:47 RealBadAngel http://i.imgur.com/Np9h5xv.jpg
01:47 hmmmm whoa cool
01:47 hmmmm what is that?
01:47 VanessaE gah, jpg for a screenshot.  I hate you.
01:47 VanessaE :-)
01:47 RealBadAngel at vertex level i can easily rotate textures
01:47 hmmmm ah
01:48 hmmmm yes, this would be way cleaner and more efficient
01:48 hmmmm i just got a wacky idea for adding tesselation...
01:48 hmmmm i kid, i kid.
01:49 RealBadAngel hehe
01:50 RealBadAngel i think facedir code will be splitted into two parts, one at vertex level responsible for rotating textures
01:51 RealBadAngel and 2nd as before responsible for picking faces from lookup table
01:52 RealBadAngel i will think about it for a while
02:37 Muadtralk joined #minetest-dev
03:20 Muadtralk left #minetest-dev
03:39 ShadowNinja joined #minetest-dev
05:16 ffoxin_ joined #minetest-dev
06:17 iqualfragile joined #minetest-dev
06:21 ffoxin_ joined #minetest-dev
06:21 ffoxin joined #minetest-dev
06:25 ffoxin_ joined #minetest-dev
06:28 ffoxin joined #minetest-dev
06:59 ffoxin joined #minetest-dev
07:39 ffoxin_work joined #minetest-dev
08:16 celeron55 http://dev.minetest.net/Network_Protocol
08:17 celeron55 is that missing anything important?
08:17 celeron55 hmm, the connection and disconnection sequences should be written down probably
08:27 celeron55 now there's a bit of that too
09:16 ffoxin joined #minetest-dev
09:19 jin_xi joined #minetest-dev
09:56 proller joined #minetest-dev
10:05 darkrose joined #minetest-dev
10:26 darkrose joined #minetest-dev
10:30 Jeija joined #minetest-dev
11:17 ffoxin joined #minetest-dev
12:35 skpfreak joined #minetest-dev
12:58 jyfl987 joined #minetest-dev
13:58 troller joined #minetest-dev
14:17 hmmmm joined #minetest-dev
15:04 iqualfragile joined #minetest-dev
15:31 Joe______ joined #minetest-dev
15:33 JoeA joined #minetest-dev
15:34 Andy_ joined #minetest-dev
15:34 AndyP_ joined #minetest-dev
15:35 JoeA Hello, myself and Andy are very interested in helping with development, are there any aspects of the game that need focusing on immediately?
15:36 VanessaE JoeA: https://github.com/celeron55/minetest/issues?state=open
15:36 VanessaE start there :-)
15:36 JoeA Thanks :)
15:39 PilzAdam joined #minetest-dev
15:53 Jordach joined #minetest-dev
15:53 Jordach joined #minetest-dev
16:30 ffoxin_work_ joined #minetest-dev
16:32 Calinou joined #minetest-dev
16:36 jin_xi joined #minetest-dev
16:40 rubenwardy joined #minetest-dev
17:22 rubenwardy1 joined #minetest-dev
17:24 iqualfragile1 joined #minetest-dev
17:32 doserj joined #minetest-dev
17:32 ShadowNinja joined #minetest-dev
18:09 doserj joined #minetest-dev
18:15 PilzAdam thexyz, any ideas about the broken win build?
18:24 proller joined #minetest-dev
18:27 thexyz PilzAdam: define broken
18:28 PilzAdam https://gist.github.com/PilzAdam/4954879
18:51 doserj joined #minetest-dev
19:11 doserj joined #minetest-dev
19:27 proller joined #minetest-dev
19:29 doserj joined #minetest-dev
19:46 ffoxin joined #minetest-dev
19:53 ffoxin_entry joined #minetest-dev
19:55 proller works! http://forum.minetest.net/viewtopic.php?id=4780
20:05 doserj joined #minetest-dev
20:09 celeron55 it's very annoying to jump up steps with this new movement
20:09 celeron55 because you have to get the player up to speed and then jump at the right time
20:10 celeron55 you can't just press forward and jump because it'll jump mainly up and then you like fall on your face because it's just so dumb
20:10 celeron55 this is stupid and i hate minecraft because of it works a bit like this
20:11 celeron55 Taoki: get back the sane behavior
20:11 celeron55 i'm not going to use minetest like this
20:11 Taoki You can walk up stairs
20:12 celeron55 ...
20:12 celeron55 a game is not fun if it is not fun to move around
20:12 Taoki I don't understand what "jump up steps" means exactly
20:12 PilzAdam the problem is the 1/4 movement when in air
20:12 celeron55 i'm actually quite angry because people screw up things that work well
20:13 celeron55 it's not like i hadn't tuned it already a lot
20:13 Taoki Yeah, air movement is 1/4. Because otherwise it felt very wrong
20:13 Taoki Old behavior here can be re-enabled from minetest.conf however, if anyone wants it on their server
20:14 Taoki Air movement feels a lot better to me, I think most people can agree. Not realistic necessarily, just more right
20:14 PilzAdam if you just hold down jump, you are in air nearly all the time, and moving with 1/4 speed is annoying
20:14 celeron55 i'm not going to even argue about this
20:14 PilzAdam wich makes climbing hills a pain
20:15 Taoki Why would anyone want to accelerate in the air?
20:15 celeron55 why NOT?
20:15 Taoki It makes no sense, you walk while on the ground
20:16 Taoki At least not in a way that this should be noticeable
20:16 PilzAdam if you climb a hill with holding down jump and collide with a node all the velocity is away and you have to accelerate with 1/4 of the normal acceleration
20:16 celeron55 i seriously hope you haven't screwed up this so that i am unable to set the good settings back
20:16 celeron55 because that is what i am doing now
20:16 Taoki On the old physics it felt very bad, because you could walk on air directly. Gave the feeling physics aren't even completed. I imagined that's something anyone would want to get fixed
20:17 PilzAdam its maybe right to not accelerat in the air, but its wrong when it comes to gameplay
20:17 Taoki Just please don't change it upstream. We don't want the ugly movement again :P
20:17 PilzAdam s/we/I
20:17 Taoki If really needed, change air acceleration to 1 from 0.5, but please not more
20:18 celeron55 Taoki: i call what you have done "minecraftifying"; it has no other sane reason
20:18 Taoki celeron55: I did not do it with any minecraft reference. I just did what felt right
20:18 celeron55 yes, and i am just about to do what feels right
20:18 PilzAdam the problem is that its server side, so changing it locally wont effect playing on servers
20:19 PilzAdam movement_acceleration_air = 2 <- feels like before
20:20 * Taoki can foresee there will eventually be a split in the development this way...
20:20 Taoki PilzAdam: Yes, setting that will get the old (and horrible) air acceleration back
20:21 PilzAdam lower air accelration makes even walking _down_ a hill slower
20:21 Taoki I hope celeron55 doesn't plan to disable diving in water too. Hopefully nothing will be changed in upstream, I think everyone will agree with the current settings
20:21 PilzAdam wich really feels wrong
20:21 Taoki PilzAdam: It shouldn't, because you only accelerate slower, you don't slow down as you walk
20:21 PilzAdam if you start walking, it is slower
20:22 Taoki It only changes the fact that while you are in the air, it takes longer to reach maximum walk velicity. If you jump while on the ground, you maintain that velocity
20:22 celeron55 why does this move as fast in water as walking is?
20:22 Taoki Yes, but it's not good to start walking in air :P
20:22 Taoki Moving in water is slower than in ground, but water has friction
20:22 celeron55 not that i really care (because it isn't annoying), but that's oddish
20:23 Taoki Anyway: If you really really must increase air acceleration upstream, please don't set it higher than 1 (currently it's 0.5). I don't think a change is needed, but if one must be done that would be a good common ground I believe
20:23 celeron55 i'm changing all three accelerations back to 4 until somebody is able to show me something that is actually better (or make a vote for such with a majority agreeing)
20:23 Taoki That's a very bad decision IMO, and will ruin a lot
20:24 Taoki I suggest leaving them as is until we discuss them
20:24 celeron55 it doesn't ruin anything; it's the same as before
20:24 Taoki Yes, and before they were horrible
20:24 celeron55 you are the one possibly ruining things
20:24 Taoki This means downgrading them to feel as basic and bad as before
20:24 celeron55 i know i am getting support from hmmmm and F00 and PilzAdam and many people
20:24 doserj joined #minetest-dev
20:25 Taoki If someone supports the old physics, they probably don't support this project advancing, and want movement to stay as ugly as it was
20:25 Taoki I suport a MineTest that advances and feels better and more modern
20:26 Taoki You're basically downgrading the physics
20:26 PilzAdam make the acceleration lower so that you feel it doesnt mean better physics
20:26 Taoki If you want a vote, I suggest leaving them as they are not and voting on how to change them
20:26 Taoki PilzAdam: What do you mean? I did, but c55 wants to downgrade
20:27 celeron55 i'm basically hitting my head in the wall because i can't understand how anyone would consider the clumsy physics of Minecraft good in any way
20:28 F00 Taoki: That's kind of disingenuous.
20:28 PilzAdam your movement feels more like RL physics because you have lowered the acceleration so that you really notice the accelration time
20:28 F00 I'm all for fast move in water/climbing/etc.
20:28 Taoki celeron55: I am because I can't understand how anyone would want movement to work like that
20:28 Taoki celeron55: Also, defaults need to be changed in 2 files, if you really insist to. Check where the same values are initialized in the client.
20:28 F00 I'm not for making nearly all the existing mine shafts into lethal death traps unless they're modified to have 100 ladder blocks at the bottom
20:28 Taoki Your commit is only half of the thing
20:29 PilzAdam F00, place slabs there and hold shift
20:29 Taoki celeron55: Also, change _fast acceleration to 4 too. Using _dast will feel reaaaaally slow :)
20:29 F00 That's another "bug" that will be fixed.
20:29 celeron55 Taoki: maybe you could tell where
20:29 Taoki **fast
20:29 F00 What I object to is
20:29 Taoki Fine, I'll help you ruin the changes I made, because I'm such a nice person. Hold on I'll look
20:30 F00 In 0.4.4, you can make a shaft into an elevator that stops at certain spots just by digging a hole and placing ladders at strategic spots
20:30 F00 I do not accept sneak/holding fast as a substitute for that because it requires player interaction
20:30 F00 Which is tedious if it's a very large shaft on a laggy server
20:30 PilzAdam grabbing a ladder after a 100+ m fall will simply break your arms
20:30 Taoki celeron55: player.cpp line 61. Blast them there too :P
20:31 F00 Yes, and carrying 3200 m^3 of cobblestone around in real life will crush you to death
20:31 Taoki PilzAdam: In reality it would. In MineTest, we must have magical physics
20:31 Taoki Cuz some people can't agree with MT becoming more modern and using saner physics like most games have
20:31 PilzAdam if nodes would support friction it would be possible to make ladders instant stop for the player
20:32 Taoki Yeah, more insane physics :P
20:32 F00 That seems circular, though.
20:32 F00 It's quite strange to have water/climbing instantly remove all your vertical inertia
20:32 F00 But while that's weird, I prefer it to sliding down 100 blocks of ladders after a 100-block fall.
20:33 PilzAdam I actually dont like this kind of glitches (ladders stopping fall) and Im happy that they are fixed
20:33 Taoki celeron55: Could you at least test air acceleration 1 or damn... even 1.5? Or in case you set ground acceleration to 4 too (also very bad and I don't think THAT one was any harm), try 2 for air, or 2.5 or something like that.
20:33 F00 Eh. I see them as bugs that have been adopted and deemed desirable by most of the community.
20:34 F00 On most servers I've visited, sneak elevators are the norm and they have a bucket worth of water at the bottom or a ladder/rope
20:34 PilzAdam I dont like a Minetest like this
20:35 Taoki celeron55: What I mean is, if you don't have anything else to work on now, maybe spend some time playing with these settings. Maybe you can find something you agree on but which doesn't make movement feel as basic and static. I don't think ground acceleration 2 was bad (it feels almost no different), maybe try 1.5 or so for air
20:35 Taoki celeron55: Also a friendly hint: Don't set fast acceleration to 4, because then accelerating and decelerating in _fast mode will feel slow. And from what I see you want all accelerations to be fast.
20:36 F00 As I mentioned yesterday, since the walk accel/decel was multiplied by BS twice, it actually used to be 40, not 4.
20:36 Taoki F00: Yeah, all settings are * BS (not sure what that means exactly though)
20:38 ShadowNinja joined #minetest-dev
20:38 F00 The motivation behind it is explained in constants.h:65
20:38 F00 But anyhow, it's 10 and its accidental double usage is why accel/decel in 0.4.4 and earlier releases is 40, not 4
20:38 F00 Which is why a value an order of magnitude less feels so tiny
20:38 Taoki celeron55: Anyway, let me know if you have more questions about how these work. Didn't mean to get pissed at you changing them, but I just ask to please test them a bit than just setting everything the old way. Maybe you can find something nice that you like as well
20:39 Taoki The point of lower air acceleration isn't to make movement in the air harder. But to remove that feeling that you are walking on air like on the ground, cuz it feels very wrong. If a higher air accel can prevent that feel, I'm ok with it
20:39 Taoki Also, ground accel 2 should really be ok and not bother anyone
20:40 PilzAdam celeron55, also change the settings in minetest.conf.example
20:40 F00 My thinking is this: Nobody will disagree that fast move in liquids and when climbing is a good thing. In fact, I bet it will cause ladders/ropes to replace sneak elevators in most instances.
20:40 Taoki Leave them there at least :P
20:40 F00 However, most of the other parts of the patch come down to taste.
20:41 F00 I personally prefer high air control and massively fast accel/decel, which apparently feels cartoony and wrong to others
20:41 Taoki F00: Best thing then is to find the most common settings for everyone. I don't want hugely slow acceleration either, but just enough so you don't feel like your speed is set directly when you move
20:41 Taoki As in, have the player feel he IS accelerating at all
20:42 Taoki An acceleration of 4 does not show this. Before I looked at the code, I didn't even KNOW horizontal acceleration existed. I thought pressing forward just moves you in that direction over a fixed pattern
20:42 F00 Well, yeah
20:42 PilzAdam Im fine with: default = 3; air = 2; fast = 20
20:42 F00 Because it's 40/s, you hit 4/s in 0.1s
20:43 Taoki PilzAdam: Sounds more ok to me than setting them all to 4, yes.
20:43 Taoki Though for fast, something like 8 or 10 is good, so you still feel a little bit of acceleration there too
20:43 Taoki Again, just enough so you can feel you ARE accelerating at all
20:43 PilzAdam fast should be definetly be that high, because its God mode and God doesnt need acceleration
20:44 Taoki Well, ok. I don't care about acceleration with _fast that much. Though I'd still encourage leaving just a little bit so there is some smoothing for a good feel
20:44 Taoki Even if you only feel it for 5 or so frames
20:45 Taoki Try 10 or 15 too or such
20:45 F00 One of the big things I object to is the different decel rate between normal and fast move
20:45 F00 Although I know why that happens, code-wise
20:45 Taoki F00: Yes, can't be helped. Unless we also remove acceleration entirely in non-fast mode
20:45 F00 Or have deceleration be a constant except when fast and fly are enabled.
20:46 PilzAdam fast = 10 is ok too
20:46 Taoki PilzAdam: Sounds great to me then
20:46 PilzAdam so: (3,2,10) is my suggestion
20:46 F00 I don't much care for accelerating for multiple seconds.
20:46 PilzAdam ^ celeron55 (if you still care)
20:46 Taoki celeron55: Can you try the values PilzAdam suggested? default 3m air 2, fast 10?
20:46 F00 Consider the case of traversing a hill
20:46 F00 Where you repeatedly whack into rising terrain and need to re-accel
20:47 Taoki I'd still say at most 1.5 fir air, but 2 is certainly better than 4
20:47 F00 *40
20:48 PilzAdam my suggestion doesnt really differ from the former values except a bit of the feeling of acceleration
20:48 Taoki yeah, that feeling is what I mean
20:49 Taoki Anyway, at least the code is there, so anyone can get the old (or new) settings on their server. Making them configurable was one point, and this proves useful
20:49 PilzAdam most people will use default values
20:49 Taoki yeah. Defaults should be somewhere in the middle
20:49 celeron55 in other news: https://github.com/celeron55/minetest/commits/menubackground
20:50 Taoki eg: I say 0.5, c55 says 4, you say 2. So something around 2 sounds reasonable for everyone
20:50 celeron55 do people approve this one
20:51 Taoki celeron55: That is some good news at least. Any screenshots? Textured menus sound really nice
20:51 celeron55 ehm... that is not what it is
20:51 Taoki <sarcasm>Let's ust make sure the texture isn't too pixelated, else we'd be copying MineCraft</sarcasm>
20:51 Taoki Ah..
20:52 Taoki Says something about menu background images
20:52 PilzAdam celeron55, I dont like the filter
20:52 Taoki Any screenshots tho?
20:52 PilzAdam just take the background of the worlds game that is selected in the world list
20:53 celeron55 i'd kind of like it if games could "brand" themselves better, in addition to having a better background images
20:53 celeron55 -a
20:53 PilzAdam I mean, I *really* dont like the filter
20:54 proller liquid want to merge! https://github.com/celeron55/minetest/pull/482
20:54 PilzAdam seperating the games isnt the right way
20:54 celeron55 PilzAdam: ack
20:54 * celeron55 is awaiting more opinions until deciding on anything
20:54 RealBadAngel http://i.imgur.com/pghAGlG.png
20:54 PilzAdam (sorting the full list would be ok)
20:55 Taoki celeron55: Oh, now I see what it does. Different background for each MineTest game. That's a perfect idea!
20:55 RealBadAngel hi all. menu background were possible before, but hard to scale
20:55 Taoki Default background might be too pixelated, but the idea and system are excellent
20:55 RealBadAngel ive made this months ago
20:56 Taoki Gonna test various physics and suggest them for now, see if I can find anything good enough
20:57 RealBadAngel celeron55, im thinkin now where to put the code for rotating tiles (withot transform stuff added to texture name)
20:57 celeron55 http://ompldr.org/vaGg2dQ/tscrot-2013-02-15_22-56-54.png
20:57 celeron55 that should explain it somewhat
20:57 Taoki Hmm... looks good
20:58 celeron55 http://ompldr.org/vaGg2dw/tscrot-2013-02-15_22-57-56.png
20:58 celeron55 i'm not sure about the world list filtering though
20:58 Taoki I like
20:59 PilzAdam Id suggest this as a background for default (warning, large): https://raw.github.com/PilzAdam/minetest/816594ac57dadcc53412c3f4185a5ad46d2344ea/textures/base/pack/menubg.png
21:00 celeron55 isn't that quite dull?
21:00 PilzAdam well, its a background
21:00 Taoki celeron55: Regarding the accelerations, I have one more wish: Please test the values PilzAdam suggested (default = 3, air = 2, fast = 10). I tried them out, and it removes that ugly feeling with air acceleration, while keeing it controlled well enough. If you can put those upstream, I won't have reasons to complain about the default physics again... they're good for me
21:00 celeron55 this isn't call of duty or mafia 2 anyways
21:01 PilzAdam I will take it for PilzTest if you dont like it
21:01 celeron55 (i'm always wondering why new games lack all color)
21:01 PilzAdam its not that I insist on using this
21:01 RealBadAngel http://i.imgur.com/aGL96cQ.png
21:02 RealBadAngel i made quite nice wallpaper a while ago
21:02 PilzAdam RealBadAngel, that is called "Fake"
21:02 Taoki RealBadAngel: That's a MineTest screenshot from the fairly close future ;)
21:02 celeron55 PilzAdam: a final image isn't something i want to choose now; i just made that as a prototype for the idea so that we can know where to aim for
21:02 PilzAdam sure
21:02 RealBadAngel PilzAdam, this is called "Wallpaper"
21:02 RealBadAngel :P
21:03 PilzAdam btw: Terasology does look like this
21:03 PilzAdam (a bit more cartoon-ish)
21:03 Taoki celeron55: I'm not sure if the branch does this already. But what I'd consider for each world is taking a screenshot before the player exists, and using that as the background next time you select it
21:03 RealBadAngel no, Minetest looks like this :P
21:03 celeron55 maybe i should make a game where you can't see anything from all the blur and lense flare
21:03 celeron55 everybody would be like "OMG OMG COOL GFX"
21:03 Taoki Same way you take screenshots with F12, but automatically when you quit. And make that a background for the given world
21:04 Taoki celeron55: Not THAT exaggerrated, but I'm hoping we will have that too :)
21:04 Taoki At least it couldn't be accused of copying MineCraft
21:04 F00 celeron55: Don't forget the extreme bloom
21:04 celeron55 F00: yeah i forgot that word
21:04 Taoki ^ yes yes
21:04 RealBadAngel http://i.imgur.com/eIg5Yyq.png
21:04 F00 Turning around should make your screen look like  you put a bunch of watercolour paints in a blender
21:09 RealBadAngel celeron55, shall i apply facedir here https://github.com/celeron55/minetest/blob/master/src/mapblock_mesh.cpp#L381 or maybe here with NDT_NORMAL case: https://github.com/celeron55/minetest/blob/master/src/content_mapblock.cpp#L150  ?
21:09 RealBadAngel what do you think?
21:10 celeron55 eh
21:10 celeron55 of course not in mapblock_mesh_generate_special
21:11 celeron55 getNodeVertexDirs or makeFastFace or something like that
21:11 celeron55 if somebody feels like that menu background stuff needs to go upstream, feel free to take it and do whatever to it
21:11 RealBadAngel i was thinkin bout makeFastFace too
21:11 RealBadAngel i like the idea, was asking you for it half year ago :)
21:12 RealBadAngel not to mention that game could play menu_default.ogg while in menu
21:12 RealBadAngel that would be somethin really cool
21:14 RealBadAngel or selected game/menu_default.ogg
21:16 Taoki joined #minetest-dev
21:21 ffoxin joined #minetest-dev
21:21 celeron55 Taoki: i pushed those values for now
21:21 Taoki celeron55: Was just checking them out. Thanks for considering my suggestion. If they don't go higher, I'm ok with those being the defaults
21:21 RealBadAngel celeron55, i think i found nice place to transform tile vertices: https://github.com/celeron55/minetest/blob/master/src/mapblock_mesh.cpp#L723
21:22 celeron55 Taoki: i think you should try something like such that some time after jumping it would be possible to control the player with high acceleration
21:22 RealBadAngel right before smooth lighting is applied
21:23 RealBadAngel i do have here all i need, needed tile and its vertices
21:23 Taoki Yeah, those defaults feel very good now.
21:24 Taoki Might look into ways to improve this again if I poke into the physics. But I'm making plans for other areas and features otherwise
21:24 PilzAdam Taoki, the long "sliding" effect after fast move is reduced a lot
21:25 RealBadAngel Taoki, let folks play with it for a while and then listen to suggestions they may have
21:25 Taoki PilzAdam: Normal move acceleration was increased a bit, so that's normal
21:26 Taoki RealBadAngel: Yeah. I'm ok with other suggestions, just really wouldn't have wanted the exact same old values back. Current ones are good since they should feel ok for those who liked the old movement too
21:26 Taoki As in, work close to the same way, but also give that feeling of being a bit more modern
21:26 celeron55 does someoone have some insights to this? < proller> liquid want to merge! https://github.com/celeron55/minetest/pull/482
21:26 Taoki But for now, the physics feel fine here... I think anyone else can agree. Would be good to let them be and focus on other important features
21:27 Taoki celeron55: What does it do exactly? The description ain't too clear
21:27 RealBadAngel celeron55, it works quite nice now, played for half an hour, filled some caves with water
21:27 RealBadAngel Taoki, it applies finite liquids
21:28 proller celeron55, PilzAdam looks too
21:28 Taoki What's finite liquids exactly tho?
21:28 RealBadAngel same as IRL
21:28 RealBadAngel make pump and pump out the ocean or a lake ;)
21:28 Taoki Kinda getting it now, but still not sure -exactly- what it does
21:29 Taoki Does it make water node sources get smaller as the fluid flows?
21:29 RealBadAngel when you will place water on a hill it will flow down
21:29 proller Taoki, water is not 4d marmalade, its water 8)
21:29 Taoki RealBadAngel: Doesn't that happen already?
21:29 RealBadAngel no
21:30 RealBadAngel it means after the while water will be down there
21:30 * Taoki scratches head
21:30 RealBadAngel not where you have placed it
21:30 proller but there is way ro place infinity source block, and it will floooood all around
21:30 Taoki If I place a water source on a hill, water does flow down like in MineCraft
21:30 Taoki ah. So the water -source- gets moved?
21:31 RealBadAngel not source
21:31 Taoki It creates additional water sources where water piles up then?
21:31 RealBadAngel water itself, there are no infinite sources
21:31 Taoki *flowing water piles up
21:31 Taoki Ok. I still don't get it as hard as I try :P
21:31 RealBadAngel theres water volume
21:32 Taoki Currently, if you place a water node somewhere, that will stay there as still water while generating flowing water around itself. What changes from that?
21:32 RealBadAngel when you place full water volume in one node wide hole it will fill it up
21:32 celeron55 Taoki: this is one thing that proves you are SO stuck in MC
21:32 Taoki Ugh! It's not MC, I just don't understand what the patch does :P
21:33 RealBadAngel when you increase the hole to be 2x2 wide, each will hold 1/4th of max volume
21:33 celeron55 Taoki: it is 8)
21:33 Taoki This IS one of the many things I believe can be improved compared to MC.
21:34 hmmmm indeed, i vouch
21:34 Taoki RealBadAngel: Ok, now it's making a bit of sense. So if I dig a hole and place a water node in it, then dig another hole next to it, both holes will then have what could be called a "water slab"?
21:34 RealBadAngel yes
21:34 hmmmm from what it looks like offhand, it seems that at least 70% of people hate the new physics
21:34 Taoki Finally I get it :) That's a great idea, totally up for it
21:34 hmmmm what they like is fast move in water and up ladders
21:35 RealBadAngel and i like gravity settings
21:35 Taoki hmmmm: The default values were just changed in GIT, and should not give anyone a reason to hate the new physics
21:35 Taoki They feel much more like the old ones now
21:35 hmmmm i hope so..
21:35 RealBadAngel but imho this should depend  on height which could be used in near core biomes, aether or space biomes
21:35 PilzAdam proller, it works quite nice compared to the version I tested before
21:36 Taoki I liked the new ones, but since many people don't this is a common ground I think everyone will find fine
21:36 Taoki Anyway, if the finite water is ready, I'd love to see what upstream
21:36 Taoki But I can imagine many ways in which it could go wrong
21:36 F00 Aside from taste stuff, there are still legitimate bugs - like shooting tens blocks off the top of a ladder when fast moving up it
21:37 Taoki That's not a bug, it's just how it works... can't he helped
21:37 Taoki **be
21:39 celeron55 it is either a bug or a lack of feature
21:40 celeron55 also, as i use e for descending, it is odd to go in water with fast_move on but not holding e
21:40 celeron55 it goes fast in that case
21:40 celeron55 that at least is a bug
21:40 Taoki No, it's intentional
21:41 Taoki When aux1_descends is 1, e is used to go down. So it can't be used as the Fast button too. Therefore, if fast_move is enabled, fast must be automatic in water and on ladders
21:41 celeron55 but when i play, i always have fast_move on
21:41 Taoki There's no logical way around that, since E takes the function of fast and makes it "go down"
21:41 Taoki yeah,
21:42 celeron55 and expect the game to behave in these ways: when i am not flying, the player goes fast when i press e, and when i fly, the player goes always fast
21:42 celeron55 in that sense, going in water is equal to walking on ground and not equal to going in air
21:42 Taoki That part should work fine
21:42 celeron55 it should descend from shift when in water, even with e being descend when flying
21:42 celeron55 and go fast by 'e'
21:43 celeron55 i belive this is how everyone else using aux1_descends wants the game to work
21:43 Taoki That part can be fixed. Although that would be confusing... you descend with E in air but with Shift on ladders and in water
21:43 Taoki It is however something I can fix, yes
21:44 Taoki So aux1_descends only affects free_move
21:44 celeron55 i don't know, but going automatically fast in water when not flying is just wrong
21:44 Taoki But I think a lot of people will find that confusing
21:44 Taoki My suggestion would be not having aux1_descends at all. But a lot of people are used to that, so they wouldn't like it
21:44 F00 Maybe the backstory is that the player character is half human, half speedboat
21:44 Taoki Using shift to go down anytime is logical
21:44 Taoki :)
21:45 proller also new masterserver with auto announce ready for test https://github.com/proller/minetest/tree/masterserver
21:45 Taoki But yeah, we should find a good way to fix this problem. Automatically going fast in water is wrong, but aux1_descends takes the E button's function
21:46 celeron55 well, going fast down could be e+shift 8)
21:46 Taoki proller: Sweet!
21:46 celeron55 in ladders
21:46 Taoki hmmm...
21:46 Taoki celeron55: Erm, it already is when aux1_descends is not enabled
21:46 celeron55 it's weird as hell from a programming standpoint, but it still could fit human's intuition when playing 8D
21:47 Taoki celeron55: When aux1_descends is off, you swim normally, including up and down. Holding e causes you to swim fast, and e + space or e + shift to sink or float fast too
21:47 Taoki The problem is that this option gives E 2 functions. IMO aux1_descends should be removed, but would it hurt people who are too used to it?
21:47 celeron55 of course it would, including me
21:48 Taoki ok. Only possibility then is to let shift descend in water and on ladders regardless. And aux1_descends only influences free_move then
21:48 Taoki It will be confusing for many, but best that can be done
21:49 Taoki Note that it's the same situation on ladders, now that fast_move can be used there. So again, even if you have aux1_descends, you go down with Shift on ladders. If that's ok I can get around to it
21:50 Taoki celeron55: ^ Do you agree with that happening in case of ladder movement?
21:52 celeron55 umm... e should descend in ladders, definitely
21:53 Taoki ok. Can't fix fast movement being always active on them in that case, as long as e is used to descend. So not sure what to do
21:53 celeron55 well, shift+e could then be fast descending
21:53 Taoki Third option would be to make fast_move on ladders and water impossible when aux1_descends. But then people who have this can't use fast movement there at all
21:53 celeron55 it just flips the speed modifier
21:53 Taoki That is an idea
21:54 Taoki And space + e fast climb
21:54 Taoki Erm no, that wouldn't work out
21:54 Taoki Yeah... the problem also exists for climbing
21:55 Taoki Shift would have to act as the Fast button in that case. But that would be a bit weird
21:55 proller celeron55, its possible to host masterserver at servers.minetest.net ?
21:55 celeron55 well... guess shift can be the speed modifier for that too
21:55 Taoki celeron55: What if aux1_descends simply flips the Shift and E keys entirely?
21:56 Taoki That would make most sense I think
21:56 celeron55 that doesn't work, because shift has been sneaking for a long time
21:56 Taoki ok
21:56 celeron55 the whole point of aux1_descends is to have the old controls that all old farts are used to
21:56 Taoki Then E would mean fast on ground, and Shift would mean fast in water and on ladders. Will be pretty confusing, but it's one of the few options
21:57 Taoki It's probably what I will go for then, since it's the most gentle possibility
21:58 celeron55 proller: your water thing just managed to completely jam a singleplayer server
21:58 Taoki Might get around to it tomorrow or in the next days if not tonight
21:58 celeron55 i dug a hole into a cave under a lake 8)
21:58 proller its wet 8)
21:59 Taoki BTW, another thing about the water patch: Can you make it so that if water flows into a hole, a water source slowly builds up in there?
21:59 Taoki That would make a lot of sense with your change
21:59 Taoki Also, this sounds like we might be able to remove the use of offseting water top texture by 2 pixels
21:59 Taoki Like shaders do currently
22:00 Taoki Since I understand the height of a water node will mean its quantity
22:00 proller liquid_relax = 1   -- will restore source blocks
22:00 celeron55 even just making fast vertical moving not possible would be better; currently i basically die if i go to a ladder
22:00 proller but water be not finite
22:00 Taoki ok
22:00 celeron55 8)
22:01 Taoki At the same time, if a water source flows too much, it should disappear, as the point where it flows appears. Basically get transferred there
22:01 proller but much better than now
22:01 Taoki Say you put a water source on a hill, and it flows into a hole. The water node on the hill will depleat as a new one(s) are built where the flowing takes place
22:01 sfan5 that'd make sense
22:01 Taoki If there's no hole for water to build in, then the water source disappears without filling anything
22:02 Taoki Like if it spills over a plain
22:02 sfan5 that makes less sense to me
22:02 sfan5 it shouldn't completly disappear
22:03 Taoki sfan5: It's like throwing a bucket of water. If you throw it on the ground, that water is as good as gone. If you throw it in a hole, that hole gets filled
22:03 proller in first one water node will drop down
22:03 Taoki I assume MineTest will never support dirt nodes containing water in themselves, so each node can have a wetness level
22:04 celeron55 proller: your code is very screwed up; this jams all the time
22:04 Taoki So that dirt or sand can absorb water. That would be fun but prolly VERY difficult to do
22:04 celeron55 i hope nobody accidentally merges it upstream at this state
22:04 Taoki If it's buggy I agree. The feature is still great otherwise
22:04 F00 celeron55: Why do you have your repo set up the way you do?
22:04 sfan5 Taoki: imagine a swimming pool, blocks would need a propety water_drains_into_block = <bool>
22:04 * F00 much prefers each dev to only have write access to their own repo
22:04 celeron55 i have some experience of finite water; it was the first kind of water i tried to implement in minetest
22:05 celeron55 i must say it is fucking hard to do anything that scales without basically consuming infinite resources
22:05 Taoki I can imagine
22:05 celeron55 so i don't really expect this to ever become actually usable
22:05 celeron55 if it becomes, then it's fine; but anyways
22:05 Taoki If it was easy MC would have had it too. AAAAND, we would have been copying it now so this feature would have been bad :3
22:06 Taoki So thank god MineCraft doesn't have this hehe
22:06 sfan5 ^ yep
22:06 proller celeron55, but now it is adjustable and disabled by default
22:07 sfan5 i wouldn't like finite water by default
22:07 Taoki proller: Does flowing water work well with this? eg: Does the height of flowing water adjust itself to the size of the blocks?
22:07 Taoki I would like it :)
22:07 celeron55 proller: how do you limit infinitely large operations in your code?
22:07 celeron55 if i attempt to drain a large lake, what happens?
22:07 proller celeron55, queue like in current code
22:08 hmmmm oh dear god
22:08 proller they will slowly drain
22:08 hmmmm a new way to kill everybody's cpu
22:08 celeron55 that does not sound like a good idea
22:08 celeron55 nor anything that would actually work
22:08 Taoki proller: BTW. Can your change be made so that "water source" and "flowing water" can become the same thing, and everything is a water node? Thgat could be kind of useful and more correct
22:09 Taoki So placing a water source block on the ground causes more of it to appear as it spills, as the original decreases. And of course the surfaces to connect in order to still LOOK like flowing water does now
22:09 proller i have tested it in more than new 100 worlds
22:09 celeron55 it needs something to very quickly and in a reasonably well cached way decide that some body of water is too large to be handled finitely and stick to it efficiently
22:09 Taoki But in reality, have everything be water sources / nodes
22:09 celeron55 there is no other way of making it handleable
22:09 hmmmm why are people making such huge changes and additions to minetest all of a sudden
22:09 hmmmm and someone just added in a JSON library
22:10 Taoki hmmmm: Maybe because it's gettind more devs and more attention. It's a great thing that this is happening at last :)
22:10 hmmmm this is not good
22:10 Taoki It's a very alive project... other FOSS projects are not as lucky
22:10 hmmmm it's going in every single direction possible
22:10 hmmmm if we accept everybody's contributions, we'll probably have 300 dependencies
22:10 celeron55 hmmmm: i approved json; it is a reasonable counterpart to curl - but your concern is very valid and i have been seeing it too
22:10 Taoki Contributions that add new dependencies are a different story
22:11 celeron55 hmmmm: note that the json lib can hang around in minetest's repo; it's not that bad that way
22:12 hmmmm now why do we need json exactly
22:12 Taoki That is a good question
22:12 * Taoki is still wondering how the sqlite vs leveldb story is goind too
22:12 hmmmm i mean i don't think we should reinvent the wheel, but we shouldn't add in an entire library for a single feature that only makes use of a small subset of it
22:12 F00 hmmmm: Minetest won't be complete until it's integrated an httpd and a smtpd
22:13 Taoki I agree there
22:13 F00 As well as maybe a PDF reader
22:13 Taoki I say we make MineTest an operating system!
22:13 ShadowNinja don't forget sshd!
22:14 celeron55 hmmmm: saying "subset of json" is quite... ehm... interesting, as json is so small you basically always end up using everything from it
22:14 Taoki If someone wants to add extra libs though, I have a sweet suggestion: Polyvox
22:14 Taoki (not really though :P )
22:15 Taoki Not cuz it would be bad, but it would prolly be a huge effort to switch the voxel system to Polyvix
22:15 celeron55 hmmmm: but indeed using json for one new feature and nothing else is a bit silly
22:15 Taoki **polyvox
22:16 celeron55 we could rule that minetest will do all future communication to outside things that do not use the minetest protocol by using json; that would make it reasonable-ish
22:17 celeron55 that includes in-game mod repositories and whatever
22:17 Taoki That would rock
22:17 Taoki If anyone here plays MegaGlest by chance, they have a sweet mod repository system built-in to the default menu. MineTest could surely use one too
22:18 celeron55 Taoki: it's Minetest, not MineTest
22:18 celeron55 we need a bot here to say that
22:18 Taoki right
22:18 sfan5 i can give you a bot, but not a bot that does that exact thing
22:18 Taoki BTW: Is Minetest going to be the final name? Or maybe there will be another once more of it is ready?
22:19 Taoki It does sound kind of beta-ish
22:19 Jordach FreeMines?
22:19 Taoki Would take some thinking
22:20 Taoki CeleronCraft :D
22:20 PilzAdam there are many forum threads about this topic
22:20 celeron55 Taoki: it isn't probably going to change until someone makes a succesful fork, or until something else than minetest_game gets most popular and minetest+that_one starts to get published instead of minetest+minetest_game
22:20 sfan5 i just found this: http://code.google.com/p/jslibs/
22:20 sfan5 it has nearly any library used by minetest
22:20 Taoki Makes sense
22:21 Taoki Errrm... javascript? Sounds like we'd be duplicating what we already have with LUA
22:21 celeron55 it's Lua, not LUA
22:21 Taoki The times I make that mistake xD
22:22 celeron55 taOKI
22:22 Taoki :)
22:24 hmmmm so what does taoki mean
22:24 hmmmm is that some kind of hawaiian greeting
22:24 Taoki Nothing exactly. Just a name that sorta came to mind, when I was still new to the internet
22:25 celeron55 it sounds japanese
22:25 Jordach chitchat -> #minetest
22:25 hmmmm so let me understand this exactly
22:26 celeron55 apparently it's some kind of word or name in french too
22:26 celeron55 Jordach: oops
22:26 hmmmm the point of this 'Masterserver' branch is to add name and description to the serverlist
22:26 PilzAdam Jordach, this conversation is completly relevant for development of Minetest
22:26 sfan5 minetest developers are allowed to go offtopic once a day ;)
22:27 Taoki xD
22:29 celeron55 hmmmm: it is to add a feature to the server to be able to announce itself to the public server list
22:29 hmmmm i see
22:29 hmmmm ahh
22:29 hmmmm that makes sense, i was reading the commits upside down
22:29 celeron55 json kind of makes sense in that because such web backends tend to easily support it
22:30 hmmmm dumb.. more recent events should be at the top of the page
22:30 sfan5 it would be nice if you could easily add new mod languages to minetest, we could have init.lua,init.js,init.sq,init.py,init.pl and maybe even init.[so;dll:dylib]
22:30 hmmmm yeah, now that i'm looking at it more it's making sense
22:30 hmmmm now what about the authentication
22:30 hmmmm how's that fitting in
22:30 Jordach minetest.ru could be the account holster
22:30 sfan5 *hoster
22:30 Jordach (considering the uptime is 100%)
22:31 hmmmm should be somewhere trusted
22:31 sfan5 what about gpg-signing?
22:31 hmmmm well i mean
22:31 hmmmm maybe
22:31 celeron55 hmmmm: this is pretty much copycatting openttd; i don't think it has any kind of authentication other than probably rate and ip limiting
22:31 hmmmm i just don't think it's good for rogue servers to add themselves on the list with malicious lua mods
22:32 hmmmm or exploit the clients
22:32 celeron55 the server list of openttd never seems to have problems while the game freely announces to it
22:32 hmmmm well, i don't know, can client-side lua be a security problem, or no?
22:32 hmmmm there's still the problem with the greatly increased surface area of attack
22:32 celeron55 clients don't run any lua code at the moment
22:33 hmmmm could've sworn they did.. must be going bonkers
22:33 hmmmm someone mentioned client-side lua at some point
22:33 celeron55 but a server could possibly attack using crafted images, models or exploiting some buffer overflow or similar stuff in the protocol
22:33 hmmmm right
22:33 hmmmm which is what i'm afraid of
22:33 celeron55 the attack surface is very large
22:33 ShadowNinja Jeija has a proof of concept for client-side Lua mods
22:34 celeron55 it's similar to web browsers; almost impossible to keep perfect track of
22:34 sfan5 think the other way 'round: clients exploting servers
22:34 celeron55 the server has much smaller attack surface
22:34 celeron55 it doesn't open images or models
22:34 hmmmm so like
22:34 sfan5 the protocol?
22:34 celeron55 altough it runs mods that can have exploits
22:35 celeron55 so it could be larger too...
22:35 hmmmm yeah but that needs to be done at the server admin's beheast
22:35 hmmmm what's the verdict?  do we add in something to prevent rogue servers from spamming/advertising on the announce list?
22:36 hmmmm like maybe they apply for their server to be registered
22:36 celeron55 there is nothing we can do
22:36 hmmmm and if we get reports from them we remove it ourselves
22:36 celeron55 well, except registration; but it doesn't really help either and very much is a pain in the ass for legitimate use
22:36 hmmmm if minetest gets big and lots of people use it, this really might be a problem in the future
22:37 hmmmm i can see a LandMine paying somebody to write up an exploit for libpng or whatever and then hosting this rogue server
22:38 ShadowNinja imo a registration requirement should be an option for server lists to impose
22:39 celeron55 there'll likely be mostly just one server list that everyone uses; it's not like people would care to change the default one
22:39 hmmmm i don't think there are going to be many server announcers aside from the official one, honestly..
22:40 ShadowNinja good point
22:40 ShadowNinja I think the official one should have some registration
22:40 ShadowNinja just not anything to hard to get into
22:41 iqualfragile joined #minetest-dev
22:41 celeron55 what's the point of automatic announcement then?
22:41 celeron55 and such registration doesn't keep anyone out who wants to exploit things
22:42 celeron55 he'll just register 20 accounts and switches between those
22:42 hmmmm true..
22:43 sfan5 yeah
22:45 celeron55 so i think it will either be done fully openly announceable or we will have a website with a non-automated server list (somewhat similar to current servers.minetest.ru)
22:46 Taoki BTW. Is it possible / will it be possible for different minetest games to have their own versions of minetest.conf?
22:50 celeron55 not possible; could possibly made possible if we can figure out a good way of doing it
22:51 celeron55 it definitely is possible to move settings to the game's game.conf, but some kind of a settings overlay thing needs much more consideration
22:55 Taoki Ah, so there is a game.conf already
22:56 Taoki Wondering about things like each game having its own player physics, with the new system
22:56 Taoki But that means it should be possible
23:00 celeron55 not directly; game.conf has nothing to do with the global settings
23:00 Taoki ah
23:09 proller celeron55, in luquid i can make: if queue
23:10 proller > 5000 -> force to infinite without leveling
23:13 proller or completely to old algorithm
23:15 proller most lags from mapgen, he adds to queue many blocks, but they calculated in 10-30 secondt
23:15 proller btw, you set liquid_update = 0.2 ? or 0.1
23:24 pierreghz joined #minetest-dev
23:26 celeron55 i think this is starting to be relatively good http://multa.bugs3.com/minetest/forum/
23:28 Jordach celeron55, its rather mature too
23:30 doserj joined #minetest-dev
23:31 celeron55 more actual modders should go and attempt to use it, and shout at rubenwardy about all remaining rough edeges in functionality
23:31 celeron55 (http://forum.minetest.net/viewtopic.php?id=3656)
23:32 Jordach celeron55, if it were more intergrated with the current site it'd more than likely be more useful
23:32 celeron55 it's kind of up to thexyz of when it would be hosted at an official location (because it integrates with the forum accounts)
23:33 Jordach mods.minetest.net
23:33 celeron55 also, i don't think it shouldu se such a direct copy of the main site's layout; it needs some kind of variation
23:33 Jordach some gradients to shine it up a little?
23:33 celeron55 ugh
23:34 Jordach or just different colours in general
23:35 celeron55 more like a smaller title thing and some other variation of the colors
23:36 RealBadAngel celeron55, well it works, proper tile is picked up in getNodeTile, rotation is passed to makeFastFace and here vertex_dirs are rotated.
23:36 celeron55 and a better layout otherwise; for example the box at the right makes it look horribly cramped
23:36 RealBadAngel looks like i got rid of ^[transform even for default facedir
23:37 celeron55 the information is at roughly the right place but it needs tuning
23:37 Jordach celeron55, on this 17'' monitor everything is nice and clear
23:38 Jordach and i know VanessaE complained that somethings arent in the right place
23:39 ShadowNinja joined #minetest-dev
23:41 celeron55 i kind of hate the minetest.net layout
23:41 celeron55 and style
23:41 RealBadAngel http://pastebin.com/ze6zp3vd   and    http://pastebin.com/sj7rVBNz
23:42 RealBadAngel celeron55, how about now?
23:42 celeron55 it suits reasonably as a "front door" to minetest, but it doesn't work for pages with a lot of functionality
23:42 Taoki celeron55: Kinda asked this before on the channel, but was curious what you think: What do you think about supporting svg textures? Given the simplistic style, some mods could go very nicely with vector graphics :)
23:42 Taoki But... does Irrlicht support that?
23:42 celeron55 Taoki: no and not trivial
23:42 Taoki Ok. Would have been pretty awesome
23:43 celeron55 well, not really
23:43 celeron55 it's just a matter of rasterizing them at texture load time rather than texture pack packaging time
23:43 Jordach the textures would be of infinite scale
23:43 Jordach (if left to render as an svg would.)
23:43 Taoki celeron55: I was actually thinking of rendering them as vector images, so they'd have infinite resolution. But Irrlicht would need to support that
23:43 Taoki Yes, what I was thinking about
23:43 celeron55 i don't think anything supports that
23:45 Taoki Web browsers do :P But I'm not aware of 3D engines either
23:45 Taoki Theoretically, it wouldn't be impossible. But I don't know when the irrlicht people will ever add that
23:45 celeron55 if you want to render stuff at realtime in 3D, you need to rasterize all textures in any case
23:46 Taoki Well, svg could still be used to generate png versions of textures based on screen resolution and such. But that would be sorta silly
23:47 Taoki So if someone has a small screen resolution, the engine converts them at the highers resolution they can't notice. But that would be a bad idea so no
23:47 Taoki Would only be of use if they could be rendered in realtime
23:56 RealBadAngel celeron55, can you take a look at new facedir code and comment if it is acceptable now?

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