Minetest logo

IRC log for #minetest-dev, 2021-06-06

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

All times shown according to UTC.

Time Nick Message
01:03 kilbith joined #minetest-dev
01:04 kilbith_ joined #minetest-dev
01:51 jonadab joined #minetest-dev
02:00 kilbith joined #minetest-dev
02:01 jonadab joined #minetest-dev
02:02 MTDiscord <Jonathon> sfan5 or krock: could this old PR https://github.com/minetest/minetest/pull/8448 be looked at if you have a chance? it already has one approval and would be useful to have
02:41 kilbith joined #minetest-dev
04:00 MTDiscord joined #minetest-dev
04:13 kilbith joined #minetest-dev
04:18 v-rob joined #minetest-dev
04:26 specing_ joined #minetest-dev
04:34 olliy1or joined #minetest-dev
04:38 Pexin joined #minetest-dev
04:40 celeron55 joined #minetest-dev
05:03 CeeGee joined #minetest-dev
05:46 olliy joined #minetest-dev
06:37 basxto joined #minetest-dev
06:37 luk3yx joined #minetest-dev
06:59 \c joined #minetest-dev
07:34 \c joined #minetest-dev
08:00 ShadowNinja joined #minetest-dev
09:03 calcul0n joined #minetest-dev
09:19 Fixer joined #minetest-dev
09:35 entuland joined #minetest-dev
10:31 \c joined #minetest-dev
10:31 calcul0n_ joined #minetest-dev
10:49 celeron55 ExeVirus said he'd like to be a triager on github, any opinions?
11:29 Krock would that imply issue management?
11:29 Krock if so, then yes I suppose it's okay.
11:35 olliy1or joined #minetest-dev
11:36 Krock @Jonathon I wonder how helpful this would be on "normal" stacks. If they consist of more than 1 item, you could no longer tell how many there are
11:36 Krock for sure, as long it's a tool, I don't see any issue with it
11:37 Krock * non-stackable tool, for that matter
12:09 MTDiscord <wwar> @wsor ^ didnt ping
12:34 celeron55 Krock: yes it means issue and PR management
12:36 Krock okay. seems fine.
13:19 Lone_Wolf joined #minetest-dev
13:27 \c joined #minetest-dev
13:55 MTDiscord <Jonathon> Krock: doesnt matter as the PR is providing options to modders if they wish to configure it. Simple example tho, you could hide the amount in the stack if you wanted it to be a mystery. Alternatively in creative you could use this to hide and overlay whatever info you think would be handy since to a degree you dont need stack sizes in creative
13:59 Krock "you dont need stack sizes in creative" is true, but you could still have large stacks (/giveme for example, or by collecting drops, whatsoever) without knowing how many there are
14:00 Krock but it's really up to the modder to take care of that
14:02 MTDiscord <wwar> People in creative need stacks for food for example
14:04 \c ok, stuff that. Has the UI been improved or even worked on in the last 4 years?
14:04 \c (rhetorical question)
14:08 \c And, why has my name been removed from the credits?
14:14 sfan5 ?
14:22 kilbith joined #minetest-dev
14:30 \c ?
14:31 MTDiscord <Jonathon> Krock: also color coding itemsta ks counts is on the table with that PR, it gives the modder more control while not affecting the defaults.
14:39 Krock color coding alone is not supported by the looks of it
14:40 Krock because the count_text.empty() check is done on the raw string, without processing color tags etc
14:40 MTDiscord <Jonathon> > I've tested using colorized strings (with minetest.colorize) now. The color works but somehow the text is extremely far to the right. I'll try to fix this. Edit: Fixed.
14:41 Krock I'll give this a try somewhen, but right now I believe it needs unescape_enriched() before the empty string check to automatically insert the stack size when none is provided
14:58 kilbith btw: \o == Zeno` in case people didn't know ;)
14:58 kilbith welcome back Zeno`
15:00 kilbith I mean \c
15:00 sfan5 in that case the answer is: it wasn't, it has merely moved down two sections
15:04 Krock oh. Zeno. wb
15:07 kilbith we need to pull back hmmmm to care about nrz
15:07 kilbith miss the good ol' fights
15:14 \c hmmmm won't come back
15:15 \c I doubt I'll come back either. I only ever did optimisations really and I don't have a linux box anymore
15:16 \c i did some stuff for whatsy on android
15:17 \c which kilbith and others got annoyed about
15:17 \c but *shrug*
15:18 kilbith ?
15:18 kilbith I don't remember that
15:19 \c kilbith, I think at the time you were annoyed that he was making money from minetest
15:19 kilbith who?
15:19 kilbith ah you mean MultiCraft
15:19 \c yes
15:20 kilbith that belongs to the past now
15:20 \c but it doesn't matter. All the old regulars (you, Vanessa, Shara, Krock, cheapie, etc etc etc etc) are here still
15:20 \c yeah it was a long time ago
15:21 \c if I could work out how to get linux to behave with a 4k monitor I'd be tempted to make changes and submit PRs
15:22 \c but, alas, linux high dpi seems to be even worse than windows 10!
15:22 kilbith works fine with Ubuntu 21.04
15:23 \c gtk or kde?
15:23 \c kde/qt
15:23 kilbith gtk 4 but not sure
15:24 \c i hate gtk. *sigh* I'll give linux another go soon
15:24 kilbith anyway, that's the least worst distro for 4K support
15:27 \c wsl is ok
15:28 \c in fact if the next release lives up to expecations then dual boot is gone for good
15:29 sfan5 it's not that mt doesn't compile on windows, just not as a convenient
15:29 sfan5 hi btw
15:30 Krock you could cross-compile it in WSL or use vcpkg + Visual Studio * (?)
15:30 Krock the build process should be somewhat easier now
15:34 MTDiscord <Benrob0329> \c: Most modern DEs have UI scaling nowadays, you just have to set it to be larger for HiDPI screens. GNOME, KDE, and Cinnamon all have the setting (though it looks like KDE might still have some issues with it, for some reason)
15:35 MTDiscord <exe_virus> I personally do the MSVC toolchain for minetest, works but it was roughly 80 minutes to get everything just right
15:38 MTDiscord <josiah_wi> Speaking of Windows builds, why doesn't the CI find Irrlicht after I updated the binary URL? Is it not installed to root (do I need to specify an installation path)?
15:40 \c is krock still making windows builds?
15:41 \c sfan used to make win build as well, didn't he?
15:45 sfan5 @josiah_wi there is no root folder so to speak, if you look further down the file it used to specify the path to irrlicht manually
15:47 nrz_ haha kilbith, you are in a nostlagia era ? :D
15:48 Krock \c: no, since 0.5.0-dev and the move to full-featured C++11 I dropped Windows (XP). GitLab pipelines and sfan still produce them tho
15:49 \c yikes
15:51 pgimeno chit-chat to #minetest please
15:51 pgimeno (maybe a bit late)
15:52 \c pgimeno If I wasn't a core dev I'd call this chit-chat. But, I am a core dev so it's not chit-chat
15:54 \c despite the "credits" I still have commit privs to the repo
15:55 sfan5 (most people on that list still do)
15:57 \c yes
15:58 \c confirm :)
15:58 \c On discord I am Clexane
15:59 \c VanessaE support me here, lol
16:00 * VanessaE hides :P
16:00 MTDiscord <Clexane> Thanks @Benrob0329
16:00 \c cheers Benrob
16:00 MTDiscord <Benrob0329> NP
16:00 \c don't hide VE :P
16:01 MTDiscord <Clexane> I suppose I need to clone and do some dev now haha
16:01 MTDiscord <VanessaE> yep, you've been roped in ?
16:01 MTDiscord <Clexane> looks like it
16:01 MTDiscord <Clexane> well, done it once, can do it again
16:05 MTDiscord <Benrob0329> There's a lot of PR review needed right now, some good ones are ageing
16:05 \c Cloning mt now
16:06 \c tentatively my plan is to do a quick code review (meh) and profile it on linux (not windows)
16:06 \c are we still on github?
16:06 sfan5 nrz_: are you happy with 11244 now?
16:06 sfan5 \c: yes
16:07 \c ok, well I'm bored and VE has convinced me to start coding again
16:08 \c so... let's do it
16:08 \c VE didn't really convince me (or ask me) I just enjoy teasing her
16:13 nrz_ sfan5: go ahead
16:13 VanessaE :P
16:17 VanessaE all I had to do was say "good to see you're still around :)" and after some catching up, he's all "welp, that's it, I'm all-in".... so my mere presence inspired him  :)
16:20 \c last time my focus was on optimisation
16:20 \c this time my interest is in the UI
16:21 \c just gotta get this dev env set up
16:22 \c I'll probably profile and do performance PRs again, but I really think the UI can do with some work
16:22 rubenwardy \c: formspecs have had a lot of improvements, including styling, new layout elements, and others
16:22 rubenwardy V-rob is working on a formspec replacement
16:23 \c rubenwardy yeah, I'll check that out
16:23 rubenwardy There's designs for a new mainmenu, but it's largely halted for now
16:23 rubenwardy Due to my time and the formspec replacement
16:23 \c time is always the problem
16:23 \c but.. let me install this and look where things are at
16:24 \c I don't have realbadangle or hmmm to commit >2000 line changes now, but let's just see
16:25 rubenwardy Also, inactive/previous core devs shouldn't have commit access but it's easy to reinstate
16:25 \c I just ran the pre-compiled MT on Win10 and it does seem to play nice with 4k. Nice
16:26 rubenwardy That's a surprise
16:26 rubenwardy :D
16:26 \c lol, maybe. But it does look ok
16:26 \c (as far as scaling goes anyway)
16:26 specing joined #minetest-dev
16:27 Pexin 5.4.1?
16:31 sfan5 windows used to not do dpi scaling, it was just missing reading of the system dpi and I added that a while ago
16:34 sfan5 merging #11244 in 15m
16:34 ShadowBot https://github.com/minetest/minetest/issues/11244 -- Shadow mapping render pass by 0xLiso
16:35 VanessaE noooooooo
16:36 VanessaE it's still too buggy
16:36 MTDiscord <Benrob0329> Its optional though, and can see incremental improvements with wider testing
16:37 VanessaE I thought devs were rather opposed to that?
16:37 VanessaE I mean the "merge if buggy, fix later" idea
16:37 sfan5 no matter what the solution is keeping the 100+ comment PR around is not feasible
16:38 VanessaE mmh
16:38 sfan5 and I have confirmed that it at least doesn't break anything if disabled ;)
16:38 VanessaE fine, but that just means effort needs to be redoubled to tune it up before next MT release
16:39 sfan5 we have the time
16:50 Pexin the project has the time, but nobody working on the project has the time  :]
17:05 kilbith joined #minetest-dev
17:07 MTDiscord <Jonathon> lol, yes
17:08 kilbith now we need more emphasis on zstd and Lua async jobs and we are good to go with the greatest release of MT
17:13 sfan5 eh, the engine could also use some de-hardcoding and APIs to control stuff
17:14 sfan5 for example the camera, then someone could make an RTS using the engine
17:15 sfan5 merging game#2884, game#2872 in a variable amount of minutes
17:15 ShadowBot https://github.com/minetest/minetest_game/issues/2884 -- Fix problem with sharing table in door nodes lua definitions. Issue 2882. by sfence
17:15 ShadowBot https://github.com/minetest/minetest_game/issues/2872 -- Filename Incorrectly Documented by AntumDeluge
17:25 kilbith joined #minetest-dev
17:43 kilbith joined #minetest-dev
18:11 tekakutli joined #minetest-dev
18:29 kilbith_ joined #minetest-dev
18:33 jonadab joined #minetest-dev
18:53 kilbith joined #minetest-dev
19:21 CeeGee joined #minetest-dev
19:23 kilbith_ joined #minetest-dev
19:51 kilbith joined #minetest-dev
19:56 nrz_ sfan5, we can invent a camera API creating the client stuff yeah :D
19:56 nrz_ that can permit for example to use camera from vehicles instead of char :D
19:58 MTDiscord <exe_virus> Wouldn't it be better to make the character model a vehicle, and allow more varied player phsyics, so that you could have a lag free vehicle?
19:58 MTDiscord <exe_virus> That's my thinking...
19:58 sfan5 we can implement client-side prediction in addition to that
20:02 MTDiscord <exe_virus> true, for the most part. The use case I'm thinking of is a boat, or bike, or car, etc. The client won't be able to predict changes in that movement besides previous acceleration. For example, If I hit the down key on a car, bike, or boat, I would slow down the next frame, but according to my acceleration it wouldn't change and the client would be out of sync. If instead I was in a plane with inverted controls, it would instead raise
20:02 MTDiscord my nose upwards. It's quite hard to build that kind of prediction into the client, unless it's scripted?
20:06 sfan5 scripted prediction is what I mean
20:32 MTDiscord <exe_virus> okay, on the same page. thanks!
20:55 MTDiscord <exe_virus> So discussing with benrob just now, it occurred to us that minetest should support a newer format for animated meshes, instead of .b3d and no longer blender supported .x files. The one we chose as a first go is .gltf from Khronos Group. I found a MIT license parser and loader for the format, .gtlf supports a single binary file, containing textures model, etc. https://github.com/syoyo/tinygltf  I personally think we should only
20:55 MTDiscord support the separated one that allows us to use our current texture loader and only parse the .gltf and .bin file (animations, materials / verticies, normals, texture coords ). Thoughts? should I just make an issue for the subject?
20:58 sfan5 agree on only loading model (or related) but not texture data from gltf
20:59 sfan5 I'm pretty sure we're better off writing our own GLTF loader instead of integrating a third-party one, we already have a json parser and the rest should just be some binary wrangling
20:59 kilbith joined #minetest-dev
21:00 MTDiscord <exe_virus> well, they are using the same json library we are, so I figure we can steal their parser code
21:06 MTDiscord <exe_virus> It's MIT so we give credit for the base code and move on. Mostly I appreciate the structures for parsing are already written and all possible options are dealt with.
21:11 kilbith_ joined #minetest-dev

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