Time  Nick      Message
10:05 sfan5     merging #14954, #15179, #15165, #15177 in 12m
10:05 ShadowBot https://github.com/minetest/minetest/issues/14954 -- [no sq] Bulk LBMs by sfan5
10:05 ShadowBot https://github.com/minetest/minetest/issues/15179 -- Fix TODO by wrrrzr
10:05 ShadowBot https://github.com/minetest/minetest/issues/15165 -- Drop fixed pipeline lighting stuff by grorp
10:05 ShadowBot https://github.com/minetest/minetest/issues/15177 -- [no sq] irr_ptr improvements, and use it more by Desour
10:06 Desour    please wait with #14954 a bit. I have some comments
10:06 ShadowBot https://github.com/minetest/minetest/issues/14954 -- [no sq] Bulk LBMs by sfan5
10:08 sfan5     okay
10:08 sfan5     merging everything else thne
10:08 sfan5     then*
11:02 Desour    why does CIrrDeviceOSX.mm have the file suffix .mm, and not .cpp?
11:05 Desour    a nvm, it's objective C++, and just *looks* like c++
11:05 TenPlus1  Hi folks
11:06 TenPlus1  https://github.com/minetest/minetest/issues/15073
11:06 Desour    !title
11:06 ShadowBot [verticalframe adjustment · Issue #15073 · minetest/minetest · GitHub
11:07 TenPlus1  hi sfence
11:10 TenPlus1  turns out the [sheet texture modifier can be used to do the same thing [verticalframe does
11:10 TenPlus1  so wouldn't it be wise to tweak [verticalframe to work simpler before it's widely used ?
11:12 TenPlus1  if the engine already knows the width of an image provided, [verticalframe could be simplified to where you only need to give it an actual number which returns that frame
11:12 Desour    you mean the height, right?
11:12 TenPlus1  nope, width... if engine knows width it can determine the frame height automatically and know how many frames it has to deal with in an image
11:12 Desour    better would be aspect ratio, to be resolution independent
11:13 TenPlus1  e.g. "animated_image.png^[verticalframe:1" or 5 or 100, error correcting itself to the last frame
11:14 TenPlus1  this would compensate for texture packs changing the size of a texture, or even the number of frames in an animation
11:15 Desour    "animated_image.png^[verticalframe:1" doesn't include a width
11:15 TenPlus1  doesnt need one, the engine knows the width of an image already and can supply it in the calculations
11:16 TenPlus1  if you have to supply dimensions then [sheet does the same thing, so why have [verticalframe to begin with
11:18 TenPlus1  The problem is that "fire_basic_flame_animated^[sheet:1x8:0,0" and "fire_basic_flame_animated^[verticalframe:8:1" do exactly the same thing
11:18 TenPlus1  so while it's still a new feature we can change [verticalframe to work better in our favour
11:21 Desour    it's not a new feature
11:22 TenPlus1  I assumed it was added in the last few versions
11:22 Desour    and idk how a single frame number should be enough for the engine to know which part of the image to use
11:23 TenPlus1  if "[verticalframe:8:1" is enough for it to know 8 frames, give me frame 1. it knows the image dimensions already
11:23 TenPlus1  so changing it to "[verticalframe:1" would easily know the width of an image x frame number for height
11:24 Desour    oh, do you mean we should use the dimensions of the underlying image, from the overload ^ operator?
11:25 Desour    or where are the dimensions coming from?
11:25 TenPlus1  the engine knows what the image dimensions of say "fire_basic_flame_animated.png" will be, so width * frame number gives location of that frame
11:25 TenPlus1  and if the frame number it larger than the actual image, say 100, it defaults to the last frame available in that image
11:26 Desour    but images aren't quadratic in general
11:26 Desour    square*
11:27 TenPlus1  when dealing with nodes arent all animated image frames the same height as width ?
11:29 Desour    if you put a rectangular image on a square face, it gets stretched. that's a feature, mhm
11:29 TenPlus1  dang it! in that case it's a no-go for the changes
11:30 Desour    anyway, so what you're suggesting is basically a verticalframe which assumes square frames?
11:30 TenPlus1  yeah
11:30 TenPlus1  since both modifiers do the exact same thing, just different values
11:30 TenPlus1  seemed a waste
11:31 Desour    I'd generalize that to a modifier that takes an aspect ratio parameter. yours would then be the case where that parameter is 1
11:32 TenPlus1  that woudl help for sure, especially when textures change size in a texture pack
11:33 TenPlus1  or even a command that returns the dimensions of an image loaded
11:33 Desour    the server doesn't load images
11:33 Desour    and it doesn't have the player's texture pack
11:34 TenPlus1  fiddly problem, lol
11:35 TenPlus1  at least talking it over means I can close the issue I opened since it's not viable
11:35 Desour    yeah, it's not new, we can't just change its behaviour x)
11:38 TenPlus1  thanks for the assist :P
15:29 sfan5     rubenwardy: ping for android build
15:30 sfan5     i dont even know who else has the signing password
15:36 [MatrxMT] <Zughy> any idea why the server is not starting? We've just updated to 5.9.1 https://i.imgur.com/ECcsP7O.png
16:42 [MatrxMT] <Zughy> no one?
16:45 celeron55 well there's nothing in the screenshot pointing to anything
16:45 celeron55 try --trace (but it probably won't help either)
16:46 celeron55 i'd try setting some breakpoints in a debugger and seeing if the backtrace at some point gives any hints
16:47 celeron55 (if --trace doesn't bring up anything usefu)
16:47 celeron55 +l
17:38 Krock     sfan5: n_r_z should also have access according to the wiki page
17:38 Krock     the website needs a quick update as well
18:36 [MatrxMT] <Zughy> also, I don't know if it's the 5.9.1 client but playing on a 5.8 server (the same one I used to play), now there are huge peaks of lag where HUDs sometimes arrive with even 10 seconds of delay
18:42 Krock     Zughy: try to reproduce on 5.9.0
18:43 [MatrxMT] <Zughy> on it
19:04 [MatrxMT] <Zughy> honestly can't tell, the host is making the whole server lag. Not worth to check
19:32 MTDiscord <landarvargan> Does anyone play on 5.9.1 with enable_build_where_you_stand = false (the default) ? My local client is broken or it's forced to true (What I normally play with) since somewhere between 5.8 and 5.9
19:55 Sokomine  hi. i still have that compile problem:  minetest/lib/tiniergltf/tiniergltf.hpp:3:10: fatal error: json/json.h: File or directory not found.
19:55 Sokomine  still the same problem as i had some days ago (currently limited way to deal with things)
19:58 MTDiscord <landarvargan> Did you run the dependency install in https://github.com/minetest/minetest/blob/master/doc/compiling/linux.md ? I see they mentioned you should create an issue
20:02 Sokomine  aaah! thank you! that helped. i somehow hadn't checked that current list of necessary packages :/
20:05 MTDiscord <landarvargan> Is something else broken? Or did you mean 'fixed'?
20:05 Sokomine  compiling nicely now. 70% done (slow notebook)
20:11 MTDiscord <landarvargan> It looks like having noclip enabled by fly disabled will allow you to build inside yourself regardless of the setting, and CTF must do something weird that causes it to break when placing at your feet, I can fix it by adding 0.01 to index 2 of the player collisionbox (It's 0 by default AFAIK)
21:10 MTDiscord <luatic> Sokomine: That was determined to be a bug with the build-related code I introduced, I'll try to fix it. In the meantime you can jsoncpp from your package manager as LandarVargan suggested.
23:06 [MatrxMT] <Zughy> so apparently if some dependency is not satisfied, 5.9.1 won't tell you shit and the server will simply die (see previous screenshot of mine)
23:06 [MatrxMT] <Zughy> it took us 3 hours to debug what the hell was going on
23:07 [MatrxMT] <Zughy> server is running inside docker, Alpine
23:08 [MatrxMT] <Zughy> I don't seem to recall the same thing happening with MT <5.9, so I guess (and I hope) it's a regression, because it's really bad UX
23:08 [MatrxMT] <Zughy> for the sysadmin
23:24 [MatrxMT] <Zughy> #15183
23:24 ShadowBot https://github.com/minetest/minetest/issues/15183 -- Server silently crashes if hard dependencies are not satisfied