Time Nick Message 09:54 Baytuch Greetings. I probobly fix my internet connection. I am sorry for re-connections recently. 11:32 Zughy[m] Meeting reminder, in about 7:30h 12:28 MTDiscord Android fixes: https://github.com/minetest/irrlicht/pull/128 https://github.com/minetest/minetest_android_deps/pull/2 https://github.com/minetest/minetest/pull/12700 18:17 sfan5 merging #12692, #12641, #12417 in 5m 18:17 ShadowBot https://github.com/minetest/minetest/issues/12692 -- Limit force shadow update to urgent blocks by x2048 18:18 ShadowBot https://github.com/minetest/minetest/issues/12641 -- Add handling of environment variables to control terminal/logging colors by lionkor 18:18 ShadowBot https://github.com/minetest/minetest/issues/12417 -- Docs: add "flip moon texture" into breakage file by Zughy 18:27 MTDiscord How many core devs does it take to add one line to docs? :D 18:28 celeron55 looks like github hates me. no pages will load and i can't write more hype comments to my PR 18:29 celeron55 now it works again 18:32 rubenwardy meeting in 30 minutes 18:45 celeron55 i'm sleeping in 30 minutes... good timing 18:46 celeron55 well, should be 19:03 nrz celeron55, time to stop sleeping, you will join NATO to be with us soon ? 19:10 Zughy[m] ?️ 19:24 rubenwardy !dev Meetings 19:24 ShadowBot Meetings - Minetest Developer Wiki -- http://dev.minetest.net/Meetings 19:25 rubenwardy sfan5, Krock, @x2048, HuguesRoss, nrz, nore 19:25 rubenwardy maybe we should move an hour earlier 19:27 Krock oh. totally missed it. 19:30 rubenwardy my internet connection is playing up 19:30 Krock > https://github.com/minetest/minetest/issues?q=is%3Aopen+is%3Aissue+label%3AQuestion 19:30 Krock I suppose this is to process? 19:34 Krock IMO the "Question" label should be avoided. Nobody knows when to close it 19:34 Krock they stay open forever, even if there's nothing more to discuss or answer 19:37 celeron55 these seem like one of the least productive issues to discuss 19:38 rubenwardy I think it's more "please look at these questions at some point" than discuss now 19:38 Krock > #12367 19:38 ShadowBot https://github.com/minetest/minetest/issues/12367 -- Add minetest.get_player_window_information() by rubenwardy 19:38 Krock concept question? 19:39 rubenwardy I don't think sfan5 was happy with the privacy element 19:44 Krock I don't see much of an issue, but at the same not much of an use-case 19:46 Krock although the maximal non-shrinking formspec size would be a helpful addition 19:46 rubenwardy this is one of the highest requested features from modders 19:46 rubenwardy just look at the reactions 19:47 MTDiscord privacy is simply an irrelevant argument 19:48 MTDiscord I dont mean to say that in a provocative way, but it is simply true. As a game developer and by extension application developer, having the window size is a basic and integral need for visuals 19:48 celeron55 i agree mods need a way to know the fullscreen formspec size 19:48 rubenwardy yeah seems like a non issue to me. If users complain then we can add a setting later 19:49 MTDiscord And it may feel like semantics, but its more games that need to know than mods. A game is an entire experience, where knowing the visual setup of the user can make or break the feel. 19:50 Krock I mean there might be a more helpful was for modders. formspecs: maximal dimensions. HUD: x/y position increment per "size" unit 19:50 MTDiscord Absolutely useful for mods, but it is integral for games 19:50 Krock s/was /ways / 19:51 MTDiscord Krock: Idealized solutions would be considerably more work to implement 19:51 MTDiscord Consider the simple use case of drawing a circle on the screen (e.g. sniper scope) 19:51 Krock luatic: maybe more work, but it's supposed to be usable by mods 19:52 MTDiscord Side-note, in the future will there be a client-sent window change event? Because with this PR (which is great, mind you) we'll have to do regular polling to detect changes 19:52 MTDiscord Suddenly you need to introduce vh / vw units like CSS to make this possible 19:52 MTDiscord IMO the PR should have been designed using events 19:52 MTDiscord https://github.com/minetest/minetest/pull/12367#issuecomment-1140455557 19:52 MTDiscord Well you need a server request to request the data in the first place 19:52 MTDiscord but yes, ruben answered "that can come later" 19:53 MTDiscord Ah, thank you 19:53 MTDiscord no, we might as well have the client push it when it changes 19:53 MTDiscord Well no, the client can choose to never send it 19:53 MTDiscord the client can also choose to not respond to server requests 19:53 MTDiscord You need both 19:53 MTDiscord Irrelevant anyway 19:53 MTDiscord Moot point 19:53 MTDiscord My two cents given, moving on 19:53 MTDiscord the client should send at join and then at every change IMO for minimum unnecessary network traffic 19:54 Krock I don't see much point in introducing an API now that needs additional code so that it's actually reliably usable 19:54 rubenwardy it's already reliable, just needs polling 19:54 Krock for example, you need to know the spacing of HUD elements, or their size respecively 19:54 rubenwardy the client pushes the data at join and when the data changes. There's no need for the server to request 19:54 Krock also size considerations for formspecs 19:55 rubenwardy Krock: you can already use this for fullscreen formspec . See the devtest code 19:55 MTDiscord ah, so ruben implemented it properly at the network level 19:55 MTDiscord yay 19:55 MTDiscord Krock: Since when do HUD elements have spacing? 19:55 MTDiscord Text is hard to deal with, granted 19:55 MTDiscord That smells like a tangent 19:56 MTDiscord rubenwardy: So adding the event feature would be dead simple later then? Since it already sends it upon changes? 19:56 celeron55 the lua api event could be added at any point. eg. now, in a year, or never 19:56 Krock luatic: exactly they don't come with any. that's why mods need to make sure they're far enough apart so that large scalings do not make it look bad 19:56 MTDiscord I'd add the clean Lua API now and keep the polling API internal 19:56 rubenwardy We've been arguing for better approaches for about 8 years, and yet modders have been left hanging 19:56 rubenwardy yes, it would be nice to have proper containers or sizing or whatever, but this is useful now 19:57 MTDiscord ^ 19:57 MTDiscord I have used screen information to make fullscreen formspecs. It works, and if we have that information, then we can make fullscreen formspecs. Krock: it is useful, it is something we want 19:57 rubenwardy god, this is unusuable. WTF is Virgin media doing. 11 second latency 19:57 MTDiscord I think we've diverged far from the original point 19:58 celeron55 what is the polling luatic is talking about 19:58 rubenwardy mods need to call minetest.get_player_window_information() to get the info, there's no callback 19:58 Krock MisterE: there's many ways to Rome. the one in the PR does it by letting modders calculate everything manually 19:58 celeron55 the client sends events over the network, and the lua api is effectively a local getter method on the server 19:58 rubenwardy yeah 19:59 MTDiscord yeah. Better to do it that way, and then later add helper functions, than to gicve mods too little precomputed information 19:59 Krock MisterE: yet it could be reached easier by retrieving the formspec size directly 19:59 MTDiscord what else actually needs to be discussed on the PR? 20:00 MTDiscord I do not understand how getting the formspec size lets modders know the screen size 20:00 MTDiscord but, I am now retreating from this conversation, this is a dev meeting 20:00 Krock why do you need the screen size if you have the maximal formspec size? 20:01 MTDiscord We dont currently have max form size do we? 20:01 MTDiscord If we did that would effectively be the screen size, no? 20:01 celeron55 i think the screen size is useless, mods don't draw anything using it, right? 20:01 rubenwardy you can calculate the max formspec size from the screen_size and real_gui_scaling 20:01 MTDiscord you guys are killin me 20:01 MTDiscord They would 20:01 Krock GreenXenith: defined by the maximal dimensions where it does not downscale 20:01 MTDiscord Im lost, is this an existing API feature? 20:02 MTDiscord celeron, I have made fulscreen formspecs using the screen size in the main menu api 20:02 Krock you could still add or subtract 20%, send it to fullscreen and it would look okay 20:02 rubenwardy here's how you calculate max_fs_size using my PR: https://github.com/minetest/minetest/pull/12367#issuecomment-1133929847 20:02 MTDiscord Looks ok ~= full control ? 20:02 rubenwardy I don't want to document this yet as the client-side code is still a bit buggy 20:02 celeron55 can someone explain what is the thing mods couldn't do if they only got the maximum non-scaling formspec size? 20:03 MTDiscord Calculations with HUDs 20:03 celeron55 not the screen size 20:03 celeron55 so what is the unit huds use? 20:03 MTDiscord You asked what they couldnt do with only form size. The answer is HUD calculations. 20:03 MTDiscord We really need both 20:03 MTDiscord both maximal formspec size and screen size 20:03 celeron55 do huda use pixels? 20:03 celeron55 huds* 20:03 MTDiscord yes 20:04 MTDiscord I believe so 20:04 rubenwardy huds use scaled pixels, so pixel * real_hud_scaling 20:04 rubenwardy but I imagine there's hud elements that render pixels directly or something 20:05 MTDiscord Ill be blunt: Window information is a basic part of applications. It's not up to anyone to decide if it is "useful" for game developers. It is an important feature to have. 20:05 MTDiscord and besides that, it is useful 20:05 MTDiscord well, would be if we had it 20:05 MTDiscord What else needs discussing on the PR to get it merged? 20:07 MTDiscord Modders who know what they need, to do what they want, have repeatedly asked for this feature over many years. Please, dont say that they dont know what they need. 20:07 celeron55 as long as someone does a code quality review and adds the necessary formulas to docs that modders need to know i think it's fine 20:07 MTDiscord Fantastic 20:07 MTDiscord claps 20:08 celeron55 the network protocol looks ok to me, that's what i really only care 20:08 celeron55 +about 20:10 MTDiscord Next item then? 20:14 MTDiscord Guess that lived long 20:15 rubenwardy next item is https://github.com/minetest/minetest/pull/12480 20:16 rubenwardy how important is keyboard navigation support, and what should be done to improve it 20:16 rubenwardy formspecs are lacking generally in keyboard support 20:16 rubenwardy !title 20:16 MTDiscord Important to support but outside the scope of the PR 20:16 ShadowBot [Manual Squash] Redesign/unify mainmenu settings interface by rubenwardy · Pull Request #12480 · minetest/minetest · GitHub 20:16 MTDiscord (imo) 20:16 rubenwardy It's relevant because the old advanced settings supported keyboards very well, and someone complained in the issue 20:16 MTDiscord Oh, it did? 20:17 rubenwardy I agree that it's a larger formspec issue to be improved though, and this redesign will improve things for more users 20:17 MTDiscord Yeah I thought you meant it needed support for key-controls in forms in general 20:17 rubenwardy Yeah, the current All Settings dialog can be used 100% with a keyboard faily easily 20:18 MTDiscord Interesting 20:21 pumpkkin[m]1 hello everyone. I'm hoping to become part of this community one day, and my desire is to aid this game in gaining the popularity it deserves. 20:21 MTDiscord pumpkin, please move to #minetest-irc 20:22 rubenwardy Hey! o/ General discussion is in #minetest 20:22 MTDiscord there is a meeting occuring here 20:22 Zughy[m] frankly, the average is user is way more important than the geeky advanced one 20:23 Zughy[m] *average user 20:23 Zughy[m] the geeky can still find some way to implement an easy navigation because they know where to put their hands, the average user will just quit 20:27 rubenwardy Ok, I've gone through the questions, there's only one left: https://github.com/minetest/minetest/issues/7917 20:27 rubenwardy !title 20:27 ShadowBot minetest.register_schematic · Issue #7917 · minetest/minetest · GitHub 20:27 rubenwardy I don't know the answer to this, might be a bug or documentation issue 20:32 Zughy[m] actually two: #12684 20:32 ShadowBot https://github.com/minetest/minetest/issues/12684 -- Falling in liquids doesn't take viscosity into account for half a node 20:33 Zughy[m] basically no one knows what that register schematic function does? 20:36 rubenwardy that's new 20:36 rubenwardy *one old one left 20:38 Zughy[m] how about deprecating that function and then nuking it in 6.0? 20:40 MisterE[m] register_schematic? isnt that for making lua schematics? Im pretty sure Ive seen that used in commonly used mods 20:41 rubenwardy most use minetest.create_schematic and minetest.place_schematic 20:41 MisterE[m] please, zughy, do not go yeeting out functions that mods use :P 20:41 MisterE[m] oh, does that use lua tables? 20:41 MisterE[m] then I guess it is redundant after all? 20:42 Zughy[m] an example of the function can be seen in #10669 code snippet 20:42 ShadowBot https://github.com/minetest/minetest/issues/10669 -- minetest.place_schematic fails to update surrounding liquids 20:43 rubenwardy register_schematic is used by 6 packages: https://content.minetest.net/zipgrep/0512a97e-e61d-45fb-b289-eda3809510ff/ 20:43 rubenwardy 4 unique 20:44 MisterE[m] the nether mod uses place_schematic: 20:44 MisterE[m] https://github.com/minetest-mods/nether/blob/52038017f385382fb43c9d0aa1e081e6618ce1aa/portal_api.lua#L1192 20:45 MisterE[m] we cant very well deprecate these 20:45 rubenwardy we're talking about register_schematic, not place_schematic 20:46 Zughy[m] don't create_schematic and register_schematic do the same thing or I'm missing something? 20:46 rubenwardy no, create_schematic saves an area of the map to a file 20:47 rubenwardy register_schematic loads a schematic into Minetest, either by filename or definition 20:47 rubenwardy but you don't need to register, you can just place directly 20:47 rubenwardy I think register_schematic allows preloading 20:48 rubenwardy weird how it's not documented at all in the schematic section, and is instead in another section 20:52 rubenwardy I think this is a documentation issue and an unconfirmed bug 20:59 sfan5 I get the feeling todays meeting was(?) very inefficient 21:01 sfan5 anyway someone should review #12695 so essentially trivial changes don't take 2 weeks again 21:01 ShadowBot https://github.com/minetest/minetest/issues/12695 -- Cut back on Gitlab-ci & misc. updates by sfan5 21:01 sfan5 applies to some other PRs too 21:02 fluxionary_ https://github.com/minetest/minetest/blob/aa2fdc6ef6300f6b6683f96305bb1d9e63ba8ebb/src/network/connection.cpp#L1006 21:02 fluxionary_ should that `&&` be an `||`? 21:02 fluxionary_ (apropos nothing anyone is talking about) 21:04 rubenwardy sfan5: looks good to me 21:05 rubenwardy and yeah, today's meeting was pretty bad 21:06 MTDiscord offer improvements 21:07 sfan5 you need a person who takes initiative, leads through the agenda and pushes forward 21:08 MTDiscord you also need persons in general. 21:08 MTDiscord seems like not many devs were available today 21:08 MTDiscord er, consistently available* 21:10 sfan5 fluxionary_: no, definitely not 21:11 fluxionary_ sfan5: is the part that comes after the `&&` redundant then? the size will be 0, and 1 is less than 1/2 the minimum window size? 21:11 sfan5 queued_commands != queued_reliables 21:11 fluxionary_ oh i see 21:11 fluxionary_ ok, got it 21:12 sfan5 commands are things to send before they're split into packets (and assigned a seqnum) 21:12 sfan5 and since data must not be reodered you can't bypass the queue if it's not empty 21:13 fluxionary_ yeah i need to read more of this before commenting again. trying to figure out why there's a flood of "Possible packet stall" messages in the log 21:15 sfan5 I think very useful would be to attach prometheus to the queues/lists/whatever in the connection code 21:16 sfan5 then you could easily monitor for this specific case how big the queue is, the window size and how fast commands are converted to packets and how fast they end up being sent 21:16 sfan5 all in production 21:16 fluxionary_ i've suggested that to the server owner, will suggest it again 21:58 Zughy[m] RE: you need a person who takes initiative, leads etc. 21:58 Zughy[m] I'm very good at that, but since I'm not a core dev I didn't chime in 22:02 Zughy[m] I've already created the note for the next meeting, just in case 22:03 MTDiscord Aren't you a core manager or smth? 22:04 MTDiscord It makes sense that a manager leads meetings 22:04 MTDiscord At least it's how it's usually done in IT companies 22:08 Zughy[m] I still see myself as a subordinate. But if core devs are fine and I'm available in those days (like today), I'll be more than glad