Minetest logo

IRC log for #minetest-dev, 2015-06-15

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

All times shown according to UTC.

Time Nick Message
00:30 paramat joined #minetest-dev
00:32 paramat pushing soon https://github.com/minetest/minetest/pull/2777 -- Biome API: Enable generateBiomes to world base
00:51 AnotherBrick joined #minetest-dev
01:05 paramat now pushing 2777
01:06 nolsen .
01:06 nolsen I just updated with git pull already.
01:06 nolsen oh well.
01:11 paramat complete
01:14 nolsen I think I encounter an bug with static_spawnpoint
01:15 nolsen It's ignoring the line in minetest.conf
01:20 paramat sounds familiar
01:21 paramat there might be a bug report issue raised at github
01:22 nolsen static_spawnpoint = 488, 30, 476
01:22 nolsen instead, It spawns me at 0, 3, 0
01:22 nolsen Which was in the ground.
01:23 nolsen I had to dig a way out or people will do what they usually do, but more easily.
01:36 VanessaE nolsen: get rid of the spaces.
01:37 VanessaE make it: static_spawnpoint = 488,30,476
01:37 nolsen VanessaE: Tried it without spaces
01:41 VanessaE odd.
01:41 VanessaE usually works for me, but there is a network glitch that sometimes causes players to spawn at 0,0,0
01:54 nolsen #2788
01:54 ShadowBot https://github.com/minetest/minetest/issues/2788 -- Tab complete feature.
01:59 paramat hmmmm or anyone, how do i declare the blend noise params and seed in the scope of getBiome? see https://github.com/minetest/minetest/pull/2764#issuecomment-111679349
02:43 paramat nevermind i seem to be getting somewhere..
02:47 paramat slowly
03:04 Elronnd left #minetest-dev
03:04 paramat compiles but segfaults. im cargo culting stuff so above my level i doubt it's worth trying to figure this out myself..
03:05 paramat https://github.com/minetest/minetest/pull/2764
03:28 NakedFury joined #minetest-dev
03:55 hmmmm hmm
03:57 hmmmm paramat:  do you actually need to calculate the biome blending inside getBiome()?
03:57 hmmmm separation of concerns
03:58 hmmmm if your goal is to make the detail of heat maps and humidity maps more noisy, then perhaps you should be modifying that in heat_map and humidity_map
03:58 paramat erm i considered that and thought that was best
03:58 hmmmm my point is
03:58 paramat yes i would prefer that
03:58 hmmmm the biome manager shouldn't really know anything about the map seed
03:58 hmmmm the mapgen should know about the map seed
03:58 hmmmm all the biome manager knows about is heat, humidity, and height
03:59 paramat also it would be best for heatmap and humiditymap to include the blend variation
03:59 hmmmm yeah :/
03:59 hmmmm if you're going to modify the heat and humidity, then actually modify it
03:59 hmmmm otherwise this information wouldn't be available through the rest of the interface
04:00 hmmmm so if I were implementing this, I would personally put the modification inside of generateNoise()s
04:00 paramat ah so point poll blend noise and add the values to the heat/humidity maps
04:00 hmmmm it's the single most straightforward solution
04:00 hmmmm and it requires no modifications whatsoever to the biome manager
04:00 paramat yes i agree
04:01 hmmmm to answer your question though
04:01 paramat okay i will rework this , thanks
04:01 hmmmm you need to learn more C++ :p
04:01 paramat yeah!
04:01 hmmmm this is really cargo culting at its finest, except it doesn't work at all
04:01 hmmmm adding Mapgen *mg; to a class definition doesn't magically make it have the same information as the mapgen
04:01 paramat those were deperate acts..
04:02 hmmmm if you actually wanted to make things work that way, you'd have to initialize those variables some place before getBiome gets called
04:02 hmmmm like, add bmgr->mg = this; in makeChunk()
04:02 hmmmm and bmgr->params = &this->params;
04:03 hmmmm but I would recommend against that simply because you're actually modifying heat and humidity, and it should reflect this in the heat/humidity maps
04:03 paramat yes
04:03 paramat good i have a new direction =)
04:05 paramat i have some more changes to mgv7 noiseparams coming, details will be in the thread
04:07 paramat also i have a game PR that corrects 'is ground content' settings. wondering why trees/leaves are currently 'false' when other plantlife is 'true'
04:07 hmmmm non ononononono
04:07 hmmmm don't do that
04:07 hmmmm making leaves not part of ground content helps stop cavegen griefing
04:08 paramat many are incorrect
04:08 paramat aha i see
04:09 paramat yes i'll leave trees/leaves false then
04:09 paramat i didn't change that in the PR
04:10 paramat it's odd that mgv6 small caves don't overgenerate, we could have more intersections. perhaps your multi-stage mapgen could overgen small caves since it removes griefing
04:13 paramat i noticed MC caves seem to have more intersections and therefore more interesting
04:24 paramat also i'm adding savanna to the mgv5/v7 biome system, and will let part of rainforest grow in shallow water to create swamps
04:30 paramat left #minetest-dev
04:37 hmmmm hmmm
04:37 hmmmm i can't think of why small caves don't overgenerate
05:34 Megaf joined #minetest-dev
05:44 Hunterz joined #minetest-dev
06:03 CraigyDavi joined #minetest-dev
06:17 nore joined #minetest-dev
06:35 paramat joined #minetest-dev
06:36 paramat compiles and runs but no terrain generates, weird https://github.com/paramat/minetest/tree/noiseblend
06:39 paramat hi nore, https://github.com/minetest/minetest_game/pull/539 -- Fix is_ground_content settings for nodes. Please could you review this sometime?
06:53 paramat left #minetest-dev
07:20 NakedFury joined #minetest-dev
07:27 Puma_rc joined #minetest-dev
07:41 Dartmouth joined #minetest-dev
07:44 chchjesus_ joined #minetest-dev
07:49 Dartmouth joined #minetest-dev
07:52 kilbith joined #minetest-dev
07:55 Calinou joined #minetest-dev
08:00 Yepoleb_ joined #minetest-dev
08:43 Amaz joined #minetest-dev
09:45 cheapie joined #minetest-dev
09:50 john_cephalopoda joined #minetest-dev
10:42 john_cephalopoda joined #minetest-dev
11:12 nore joined #minetest-dev
11:14 crazyR joined #minetest-dev
11:14 crazyR joined #minetest-dev
11:27 Darcidride joined #minetest-dev
12:10 RealBadAngel joined #minetest-dev
12:42 Hunterz joined #minetest-dev
13:42 celeron55 https://www.youtube.com/watch?v=je-KroYSzrU
13:43 celeron55 lol i'm cringing hard at how much this is suffering due to no moddable client-side physics or camera or anything
13:44 celeron55 every core developer should watch this
14:03 Hunterz joined #minetest-dev
14:05 LittleJoe joined #minetest-dev
14:10 Lunatrius joined #minetest-dev
14:25 Lunatrius joined #minetest-dev
14:28 CraigyDavi joined #minetest-dev
14:46 Darcidride joined #minetest-dev
14:52 Niebieski joined #minetest-dev
15:07 hmmmm joined #minetest-dev
15:28 est31 joined #minetest-dev
15:28 est31 pushing #2790
15:28 ShadowBot https://github.com/minetest/minetest/issues/2790 -- Fixed crash caused by saplings trying to grow on unknown nodes by Y-st
15:28 AnotherBrick joined #minetest-dev
15:48 RealBadAngel https://cloud.githubusercontent.com/assets/2177790/8163843/cebc3308-1385-11e5-8626-3023a865c388.png
15:49 RealBadAngel ^^ connected textures
15:51 est31 also showcasing a weakness of what kenney suggested
15:51 est31 on the edges, there is this ugly diagonal
16:00 RealBadAngel yes, triangles do cut edge pixels in half
16:03 * est31 wonders whether there is a better method
16:03 est31 interestingly, in kenney's suggestion, there is a full edge for the glass texture
16:04 RealBadAngel he propapbly just painted that
16:04 est31 yes I think so
16:04 est31 perhaps one should specify 3 images not two
16:05 est31 the first is the base one which is painted always
16:05 nore I also noticed an incompatibility in his suggestion
16:05 est31 then a second which is drawn using diagonals
16:05 nore the middle glass in the top row has a white bottom-left pixel
16:06 est31 thats what we are speaking about :)
16:06 nore while in the red & blue textures, that pixel in yellow
16:06 est31 https://cloud.githubusercontent.com/assets/8872119/8164299/62beba74-1388-11e5-8acc-03de9c522379.pn
16:06 est31 g
16:06 RealBadAngel but that method will work for his another examples
16:06 est31 and a third which is drawn last using 4 sub tiles
16:06 nore est31: I mean, not that
16:07 est31 ah I see nore
16:07 nore from his spec, pixels on diagonals should be drawn with the not-connected texture
16:07 est31 from his spec, its not observable, because there is an incompatibility as you said
16:08 nore yep, there is one on glass
16:08 est31 The method I suggest with 3 images would support both use cases
16:08 nore (the other drawings correctly match the spec though)
16:08 nore so, what would your suggestion be?
16:08 est31 <est31> the first is the base one which is painted always
16:08 est31 <est31> then a second which is drawn using diagonals
16:09 est31 <est31> and a third which is drawn last using 4 (rectangular) sub tiles
16:10 nore ah, I see what you mean
16:10 nore wouldn't you get problems with transparency though?
16:10 est31 if all have an alpha channel, why?
16:11 nore well, I mean for glass, if you always draw the texture with sides, you going to have a problem
16:11 nolsen I think I found a bug with v7, there are very few trees D:
16:12 est31 ah I realize it doesnt work
16:13 RealBadAngel est31, that rather wont work
16:14 RealBadAngel how would separate diagonals and display them with triangles?
16:14 RealBadAngel you would need to make meshes for each pixel to do that
16:16 est31 the first two steps are exactly what you are doing no?
16:19 hmmmm paramat:  if you're around, check this out:  http://vterrain.org/
16:19 hmmmm specifically http://vterrain.org/Elevation/Caves/
16:26 est31 yet another preference PR https://github.com/minetest/minetest/pull/2787
16:27 hmmmm well
16:27 hmmmm it's up to you, i don't really care
16:27 hmmmm also why is darkrose still in the core developer section
16:28 hmmmm and you aren't
16:29 MinetestForFun joined #minetest-dev
16:31 est31 the last time it was updated when I wasn't core dev yet
16:31 est31 as long as it gets updated till the next release...
16:35 RealBadAngel https://imgrush.com/G4EnJd193gZx.png
16:35 RealBadAngel looks a bit better than glass
16:36 Taoki RealBadAngel: Hi :) Did you see my last message yesterday, about the new parallax?
16:36 RealBadAngel i think not
16:36 RealBadAngel hi
16:36 Taoki [02:28:49] <Taoki> RealBadAngel: Got latest master, confirming the new parallax effect works for me (AMD / MESA). Thank you... it looks so sweet and realistic!
16:36 Taoki [02:32:54] <Taoki> Strange however... it only seems to work sometimes o.o I switched back to the window as I wrote the above line, and the effect was gone... I was getting the old parallax. Also when I quit to the main menu than back in. So something might be causing it to switch between the two randomly.
16:37 est31 RealBadAngel, now, how did you do it?
16:38 est31 because this is against my theory :)
16:38 RealBadAngel est31 those textures have common pixels on diagonals
16:38 est31 ah I see
16:38 est31 still the diagonal method
16:38 est31 cool
16:39 RealBadAngel Taoki, seems like another AMD related issue, we will have to find out what happens
16:40 RealBadAngel est31, what i have done is splitting each face into four triangles instead of 2
16:40 Taoki RealBadAngel: AMD related or MESA related. But yeah... all I know is I was getting the new effect, then switched to the IRC window, then when I alt + tab'd back to Minetest I was getting the old parallax again.
16:40 Taoki Before that, I was getting the old parallax, exited to the main menu, re-entered the world, then the new one was working.
16:41 RealBadAngel Taoki, by old parallax you mean that auto generation of heightmaps stopped working?
16:42 RealBadAngel i would like to see screenshots before and after
16:42 rubenwardy joined #minetest-dev
16:42 Taoki RealBadAngel: Possibly. I mostly mean I was getting the effect before the change.
16:42 Taoki That sounds most likely
16:44 RealBadAngel est31, theres another solution to make those textures
16:45 est31 RealBadAngel, ok 4 triangles is what thought.
16:45 RealBadAngel most easy would be a script that will  calculate all the possible combinations - five of them, taking care of those pixels on diagonals
16:45 RealBadAngel and store them as animation frames
16:46 est31 I count 12
16:46 est31 err 13
16:46 rubenwardy connected looking good :D
16:47 RealBadAngel why so much?
16:47 RealBadAngel i can see just 5 :)
16:47 est31 lemme make an ascii art
16:48 est31 oh even more
16:48 est31 RealBadAngel, you disregard roatation?
16:49 RealBadAngel imho you just need base case
16:49 john_cephalopoda joined #minetest-dev
16:49 est31 yea
16:50 est31 still I count 6 RealBadAngel
16:51 est31 http://pastebin.com/2px6z4wJ
16:51 RealBadAngel ah yes
16:51 nore est31: there is however a rotation problem (i.e. bookshelf...)
16:52 est31 yes
16:52 nore and the diagonal problem (+#/## vs +#/# )
16:52 nore (the one there is with glass with kenney's suggestion)
16:53 est31 ah I see
16:54 est31 I think we need something like a programming language
16:54 est31 input: 8 bits for the tiles around
16:54 est31 output: rules to make a texture
16:57 nore isn't it a bit overkill?
16:58 RealBadAngel for sure it is :)
17:01 rubenwardy Overkill is fun for the end user!
17:03 est31 ok something simpler:
17:04 rubenwardy MTTCPL
17:04 rubenwardy MinetestTextureConnectingProgrammingLanguage
17:04 est31 you provide a map from 8 ternary values to a tile each
17:05 est31 the values are: 1 (is there) 0 (might be there) and -1 (is not there)
17:05 est31 or rather you provide an array with tuples (8 ternary values, tile)
17:06 est31 then when it comes to rendering, the array is traversed for every tile, and the 8 ternary values are compared with the actual 8 inside the world
17:06 est31 if there is a "match", the tile is "added", if there is no, nothing happens
17:07 est31 all tiles have alpha values, so you can have transparency
17:07 est31 examples
17:07 est31 1. the glass
17:08 est31 ah, dont know how to describe them
17:09 est31 but all the examples kenney suggested would work
17:09 RealBadAngel but one problem
17:09 RealBadAngel only shaders can do that
17:10 est31 gtg
17:11 RealBadAngel john_cephalopoda, check master
17:13 RealBadAngel https://bitbucket.org/prupe/mcpatcher/wiki/CTM_Methods
17:13 RealBadAngel nore ^^
17:21 Dumbeldor joined #minetest-dev
17:23 Krock joined #minetest-dev
17:31 john_cephalopoda left #minetest-dev
17:32 proller joined #minetest-dev
17:40 err404 joined #minetest-dev
18:01 * VanessaE peeks in
18:04 est31 joined #minetest-dev
18:21 est31 the other option
18:22 est31 would be to have it pre-generated
18:23 est31 all 256 combinations
18:23 est31 like RBA said inside an animation
18:23 est31 or with additional info
18:23 est31 like a table
18:24 est31 that maps the 256 combinations to those pictures
18:24 est31 with multiple combinations being abled tomap to one picture
18:25 est31 (basically what I suggested before, but already generated)
18:25 est31 so it wont need a shader
18:25 est31 RealBadAngel ^
18:25 VanessaE sounds like a memory hog
18:26 est31 VanessaE, they would have needed to be generated either way
18:26 est31 also its not 256 * texture size unless you utilize all combinations
18:26 VanessaE ...wehich you KNOW will happen.
18:26 VanessaE which*(
18:27 est31 also I wonder how RBA did it if not with shaders
18:28 VanessaE shaders need to be reserved for special effects, and I'm not sure connected textures qualify as that
18:28 est31 agree
18:28 VanessaE well, special effects and any base effect where they'd be faster purely as an alternative.
18:29 VanessaE (i.e. lighting, which I'm pretty sure shaders were originally invented for :) )
18:31 nolsen VanessaE: It appears that spawn.conf keep setting the spawn location at 0, 0, 0. Now I'm not sure if spawn.conf is part of minetest, or is a mod.
18:31 VanessaE that'll be part of a mod.
18:32 VanessaE base minetest only uses static_spawnpoint, afaik.
18:32 nolsen Well I'm not sure what mod is using it.
18:32 VanessaE (unless something has changed recently)
18:32 Hunterz joined #minetest-dev
18:33 VanessaE probably this one, nolsen:  https://github.com/Pitriss/multispawn
18:34 nolsen Yep, It is.
18:34 nolsen Thanks
18:35 Hunterz joined #minetest-dev
18:36 VanessaE yw
18:44 Hunterz joined #minetest-dev
18:44 RealBadAngel joined #minetest-dev
18:52 Hunterz joined #minetest-dev
19:04 OldCoder joined #minetest-dev
19:24 Robert_Zenz joined #minetest-dev
19:59 MinetestForFun joined #minetest-dev
20:27 cib0 joined #minetest-dev
21:33 Dumbeldor left #minetest-dev
21:44 nolsen #2792
21:44 ShadowBot https://github.com/minetest/minetest/issues/2792 -- Ability to move items into your hotbar or chest easy with shift like Minecraft.
21:51 est31 nolsen, do you want to implement it?
21:52 nolsen I would.
21:52 nolsen If I knew how.
21:52 nolsen I haven't even finished my C++ course.
21:52 est31 when I started programming for minetest, I didnt know any c++
21:53 est31 or very tiny bits
21:53 nolsen Only C++ program I written was to fool skiddies that the program had "hacking abilities" xD
21:54 est31 you only need to modify here, not write
21:54 nolsen But where would I start?
23:02 johnnyjoy joined #minetest-dev
23:37 paramat joined #minetest-dev
23:44 jordan4ibanez joined #minetest-dev
23:45 Wayward_Tab joined #minetest-dev

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