Minetest logo

IRC log for #minetest-dev, 2020-12-13

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

All times shown according to UTC.

Time Nick Message
01:25 MTDiscord <s​rinivas> hi, is it useful/possible to add in full html + css + js support for minetest formspecs... so formspecs become easier to write, and cient side...
01:30 lhofhansl I think so, but that seem to be a major undertaking.
01:31 lhofhansl And the js part... Probably not. Another language interpreter just for this?
03:03 xerox123_ joined #minetest-dev
04:16 systwi joined #minetest-dev
05:00 MTDiscord joined #minetest-dev
06:04 hecks joined #minetest-dev
07:31 YuGiOhJCJ joined #minetest-dev
08:00 ShadowNinja joined #minetest-dev
08:17 TLuna joined #minetest-dev
08:17 YuGiOhJCJ joined #minetest-dev
10:13 proller joined #minetest-dev
10:16 MTDiscord <I​hrFussel> Isn't there already the 'style' formspec element which allows css kinda?
10:16 rubenwardy It's not really CSS, and it's not full HTML/Js
10:17 rubenwardy Full HTML/JS would probably 30x the download size of Minetest, would be crazy
10:19 MTDiscord <I​hrFussel> In what areas is Js better than Lua though?
10:20 rubenwardy Loads
10:21 rubenwardy But in this case, it's the use of JS that's important
10:21 rubenwardy JS is used in the web to manipulate the DOM and implement interactive GUIs
10:22 rubenwardy Lua can be made to do the same thing in Minetest - but the point of the HTML/JS suggestion was to use existing tech
10:23 rubenwardy The DOM is the GUI éléments in a webpage
10:23 rubenwardy Lol, English/French keyboard sabotages me
10:24 calcul0n joined #minetest-dev
10:25 MTDiscord <I​hrFussel> But our formspecs aren't even remotely close to being DOM elements I'd say
11:04 rubenwardy Which is why HTML/JS was suggested
11:05 Wuzzy joined #minetest-dev
11:52 hecks the day Minetest actually uses JS alongside Lua in one executable is the day I declare it a clown engine
11:58 MTDiscord <a​ppguru> heh
11:58 absurb joined #minetest-dev
11:59 TLuna joined #minetest-dev
11:59 MTDiscord <a​ppguru> Seeing as we are already developing our own markup (formspecs) and styles (formspec styling), we probably shouldn't move to web stuff
11:59 MTDiscord <a​ppguru> Extending that with Lua would be nice, but should probably be left to SSCSM
12:00 rubenwardy formspecs should be deprecated at some point
12:00 MTDiscord <a​ppguru> yeah
12:02 hecks that point might be SSCSM with gui
12:03 hecks if you solve hud, you also solve forms
12:03 hecks and if someone really wants to use HTML, they can make a wrapper for that
12:03 hecks until then, forms aren't that bad
12:04 calcul0n joined #minetest-dev
12:05 hecks at least not bad enough to pull in an entire web browser, and let me tell you, those things usually suck
12:06 hecks every time a game pulls in a chromium process or something as a lazy solution for GUI, it causes nothing but trouble
12:06 Fixer joined #minetest-dev
12:08 hecks in some cases it outright caused crash bugs, like with Quake Live (which was just fine before the awesomiumprocess nonsense)
12:10 Wuzzy I doubt JS will ever be added. Which is good. Fuck JS! 😀
12:10 Wuzzy btw
12:10 Wuzzy #10693
12:10 ShadowBot https://github.com/minetest/minetest/issues/10693 -- Builtin translate (2nd attempt) by Wuzzy2
12:10 hecks yeah, I kinda meant to say the same, just very politely
12:10 hecks we already have a perfectly good VM available
12:17 rubenwardy I like JS, but pulling in a web browser is crazy and Lua can already fulfill JS's purpose on the client
12:19 MTDiscord <a​ppguru> We have to develop (or use) a proper GUI library for both HUD & dialogs (formspecs)
12:19 pmp-p an html5canvas emulator + a minimal js engine would not be that bad for very complex UI
12:19 MTDiscord <a​ppguru> HTML bad
12:20 MTDiscord <a​ppguru> JS bad
12:20 MTDiscord <a​ppguru> CSS bad
12:20 MTDiscord <a​ppguru> the only reason for using these is that they are "standard"
12:20 MTDiscord <a​ppguru> Well, modern JS isn't even that bad, but not minimal either
12:22 Fractalis joined #minetest-dev
12:22 hecks I would argue with the "not that bad" part, see above, this stuff isn't very high quality
12:22 hecks Lua is a similar enough language that it shouldn't be harder to manipulate a DOM or whatever with it
12:23 hecks so the only reason you could ever have to literally use JS is to pull in JS dependencies - and the uses for this I can think of tell me that this idea shouldn't even be touched with a 100 foot stick
12:23 pmp-p hecks: sure but i doubt you'll with complete ui toolkit for canvas like zebkit/zebra in lua
12:23 pmp-p -with +find
12:24 hecks someone's gonna make one eventually
12:24 hecks I'll just roll my own
12:24 MTDiscord <a​ppguru> ^
12:25 hecks and having to audit and sandbox not one but *two* VMs handling code sent from the server stinks badly
12:25 hecks just having two competing VMs stinks enough
12:25 MTDiscord <a​ppguru> https://github.com/minetest/minetest/issues/6527
12:25 pmp-p *that* is a very good argument over "use the standards"
12:26 hecks you speak as if the available JS tools are built for quality, and not to push corporate products out the door
12:26 MTDiscord <a​ppguru> TBH I like Flutter's approach
12:26 pmp-p +1 for dart+flutter
12:26 hecks see the part about games crashing or bugging out when they start bundling this stuff - the whole point of that is cutting costs
12:26 pmp-p (i hate js)
12:27 MTDiscord <a​ppguru> We obviously can't bundle Dart & Flutter with Minetest though
12:27 MTDiscord <a​ppguru> So we'd have to take some inspiration from it and roll our own
12:27 hecks a gui toolkit is nice to have but not mandatory if you have input and rendering
12:28 MTDiscord <a​ppguru> And SSCSM
12:31 hecks to all devs: can we talk about that shadow mapping branch Liso is working on?
12:32 hecks I'd really like some kind of confirmation that it won't be merged without safety brakes on it
12:32 sfan5 you mean the server should be able to forbid the client from using it?
12:32 hecks exactly
12:33 hecks I mean, not the server - the game more specifically
12:33 hecks so not a .conf entry but an api call
12:33 sfan5 same thing in the end
12:33 MTDiscord <a​ppguru> why not allow server-sent conf entries?
12:34 hecks it would be the same thing for me, but for everyone else, a .conf flag would be inferior
12:35 hecks although okay, games can come with .conf
12:35 sfan5 work to allow games/mods to better control visual features the client has (or doesn't) is still outstanding so I'd say in this state a shadow mapping feature would be merged *without* the feature you desire
12:35 sfan5 but you're correct that as an engine Minetest should provide this
12:36 hecks well, the server can already control stuff like the skybox
12:36 hecks there are actually reasons to give it a similar API - you want to disable the sunlight shadow in caves or in other situations to save performance
12:37 hecks per-object flags are very desirable too, some things do not deserve or outright shouldn't have a shadow
12:38 hecks imagine an explosion, flame or smoke effect implemented as an objectref, you'd want that to have casting and receiving both off
12:38 MTDiscord <a​ppguru> does it even work in caves?
12:38 sfan5 what do you mean by "work"?
12:38 MTDiscord <a​ppguru> display proper shadows emitted by other light sources
12:39 sfan5 to my knowledge the shadow mapping implementation is for the sun only
12:39 MTDiscord <a​ppguru> :/
12:40 hecks yup, that's the only source of it, which is why it's a good idea to toggle it off entirely if you know a player is inside a cave or in some other place without sunlight
12:41 hecks and only the game has the knowledge to do that for sure
12:42 hecks in any case, I want at least an off switch for this
12:42 hecks there is precedent for that in the form of CSM flags
12:43 hecks and API functions to toggle stuff are also common - doing this once per player on login I can deal with
12:44 NetherEran joined #minetest-dev
12:45 hecks so, how about it, can I have some assurance that #9374 won't repeat itself and I won't have to yell at anyone?
12:45 ShadowBot https://github.com/minetest/minetest/issues/9374 -- Basic model shading by dcbrwn
12:57 sfan5 it certainly won't because none of the defaults are going to change
12:57 celeron55 i agree with hecks, any kind of shading is an artistic decision especially when coming from originally having no shading at all
12:57 hecks thank you
12:58 hecks I'd really like to just focus on making my game and not having to constantly contest features that could break it.
12:59 celeron55 i probably won't be policing this in PRs though so you'll have to quote me on that 8)
12:59 celeron55 or link them to me
12:59 hecks I've done that already :(
12:59 hecks some people just don't get it
13:01 celeron55 well of course they won't
13:02 hecks which is why I wanted some kind of assurance from devs that they won't approve features without safety brakes on them
13:03 celeron55 i actually don't even know if MT currently applies any kind of shading on models or not
13:03 hecks Since that PR I linked, it does
13:03 hecks and it was a feature without an opt-out and I kinda yelled at people over it
13:04 celeron55 that being said, people also barely notice any kind of artistic quality
13:04 hecks since hardcoded lambert shading just doesn't look good on my stuff, and it wasn't a thing when I started
13:04 celeron55 most of the time people comment that things look better when in reality it's worse
13:04 hecks what you mean to say it, most people don't have taste
13:04 celeron55 yes
13:05 hecks =]
13:06 hecks https://a.uguu.se/Bu1ue0ZCoXr2_stuff.png I'm really trying to break out of that voxel=ugly routine
13:06 hecks and I've got the skills to make that happen, just please don't pull the rug from under me
13:07 hecks celeron55: I believe you've wanted to see this kind of consistent, original content for a long time
15:42 lisac joined #minetest-dev
15:46 kilbith joined #minetest-dev
15:55 celeron55 yours indeed probably is stylistically most ambitious thing i've seen on MT
15:56 celeron55 +the
15:58 celeron55 throwing a barely dressed anime character in an MT world is ambitious enough to initially feel completely wrong
16:09 fluxflux joined #minetest-dev
16:43 hecks joined #minetest-dev
16:45 kilbith joined #minetest-dev
16:50 fluxflux joined #minetest-dev
16:54 calcul0n joined #minetest-dev
17:14 Krock thank you Irrlicht for not generating an Event when Shift is released
17:14 hecks uh oh
17:18 Krock though according to the code, it seems to depend on xorg
17:18 hecks you got something I can quickly test on windows?
17:19 Krock current master.
17:19 Krock sadly not compiled
17:19 Krock perhaps you can grab a binary from the buildbot. dunno if that's possible
17:19 Krock otherwise: https://gitlab.com/minetest/minetest/pipelines/
17:20 kilbith don't you think there are much more important stuff than fixing a platform nobody cares about there
17:21 hecks are you implying nobody cares about xorg or about windows
17:21 kilbith a small bug on an irrelevant platform moreover
17:21 hecks because both are wrong
17:21 kilbith about MacOS
17:21 Krock works on Wine... lol
17:21 Krock working around an irrlicht but unreveals another one. will revert that change
17:22 kilbith it wasn't worth patching the shift-click on this
17:31 TLuna joined #minetest-dev
17:36 Krock #10721
17:36 ShadowBot https://github.com/minetest/minetest/issues/10721 -- Revert "GUIFormSpecMenu: Shift+Click listring workaround for MacOS" by SmallJoker
17:37 sfan5 lgtm
17:38 Krock how about "Q"? Does it need docs, or is stating the solution in the issue enough?
17:42 Krock assuming latter
17:43 Krock will merge game#2795 game#2794 game#2796 and game#2758 in 15 minutes
17:43 ShadowBot https://github.com/minetest/minetest_game/issues/2795 -- Use Minetest 5.3 'minetest.is_creative_enabled' API by MoNTE48
17:43 ShadowBot https://github.com/minetest/minetest_game/issues/2794 -- Improve Russian translation (carts) by RichardTry
17:43 ShadowBot https://github.com/minetest/minetest_game/issues/2796 -- Update zh_CN translation for mtg_craftguide and dye by IFRFSX
17:43 ShadowBot https://github.com/minetest/minetest_game/issues/2758 -- Fixed some mistakes in the Indonesian translations by SpicyOnta
17:49 sfan5 Minetest should switch to using its own irrlicht fork "soon" anyway so just the latter for now
17:50 Krock alright, thanks.
17:57 MTDiscord joined #minetest-dev
17:58 Krock merging
17:59 Krock done
18:02 kilbith git reset --hard HEAD~1 wasn't enough?
18:03 Krock that's only allowed in a 5 min timespan by project rules
18:03 Krock assuming the same rule as with amended commits can be applied here
18:05 lisac joined #minetest-dev
18:39 kilbith joined #minetest-dev
18:43 calcul0n joined #minetest-dev
19:21 kilbith joined #minetest-dev
19:43 calcul0n_ joined #minetest-dev
19:48 NetherEran joined #minetest-dev
19:58 hecks joined #minetest-dev
21:06 lhofhansl joined #minetest-dev
21:08 lhofhansl joined #minetest-dev
21:17 el joined #minetest-dev
21:38 fluxflux joined #minetest-dev
21:44 TLuna joined #minetest-dev
21:49 kilbith joined #minetest-dev
21:57 TLuna joined #minetest-dev
23:30 lhofhansl joined #minetest-dev

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