Time Nick Message 00:52 fox_ hello! 00:59 DeepThgt ok so mesecons inability to handle 1 tick state changes is driivng me nuts 00:59 DeepThgt this essentially makes the detector tube useless 01:00 DeepThgt its looking like digging into trying to figure out why the delays get "stuck" with a 1 tick operation is the only path i can see to fix it 01:00 DeepThgt and i am not even sure why mesecons fails with 1 tick state changes 01:03 fox whats your fav mod? 01:03 DeepThgt surely not mesecons right now 01:04 DeepThgt no offense to the work people have done it on it whatsoever 01:05 DeepThgt just, the bug of things being stuck with 1 tick operations is annoying, and im curious as to why it was handled via the "overheating" mechanic 01:05 fox yeah... 01:07 DeepThgt from what i understand the overheat was added to keep from slowing down the server with too many operations 01:07 DeepThgt which i understand as a reason 01:08 DeepThgt but this seemes to have covered up bugs as an unintended side effect 01:08 DeepThgt and made them more rare than they would have been without overheating 01:10 DeepThgt local on_state = { 01:10 DeepThgt description = S("You hacker you"), 01:10 DeepThgt lol 02:56 DeepThgt im thinking if the pipeworks detector used globalstep instead of minetest.after it would work 02:56 DeepThgt but idk at all 04:18 * cheapie emits a digilines detector tube at DeepThgt 04:19 DeepThgt is that my real answer? 04:19 cheapie It's not an ideal answer, but it's one possible alternative. I reported the bug you're complaining about several years ago and I don't think much has changed since then :/ 04:21 cheapie I should probably fix it, but I haven't actually used that particular part in ages now so I keep forgetting it even *has* that bug. 04:25 DeepThgt a solid workaround is acceptable 04:25 DeepThgt ty very much 04:25 DeepThgt although i did learn a lot chasing that one... 04:30 DeepThgt i am actually interested still in finishing the job of hunting down its cause and fixin 04:30 DeepThgt but i still have a bit to learn i think to do so 04:31 DeepThgt ive just started digging into rubenwardys book and the api documentation deeply when trying to figure out the cause of the behavior 04:32 cheapie Another workaround that might or might not work (haven't tried it) is connecting a Luacontroller to the normal mesecons detector tube and doing things just based on the 'on' event (ignoring the 'off' and the pin state). 04:32 cheapie (but if you're going to use a LuaC you might as well just use the digilines one) 04:32 DeepThgt ya my other idea i was about to try is use a luac to replace the sticking delay 04:33 DeepThgt but ill just do your way 04:34 cheapie I have a tendency these days (when I play MT at all, which I am currently not) to just use one Luacontroller for everything (or nearly everything) and few if any mesecons. Digilines becomes quite powerful with a few extra mods using it :P 04:34 DeepThgt ya my lua skills are subpar so ive been doing it all in "hardware" 04:35 cheapie Time to work on the Lua skills :D 04:35 DeepThgt ya especially since now the issues im encountering with running the server involve modding 04:35 DeepThgt like adding intercompatibility to our modset 04:36 DeepThgt eventually fixing lwcomponents 04:36 cheapie See https://github.com/mt-mods/digistuff (full disclosure, originally one of my mods, this is mt-mods's continuing development project of it) for, among other things, a bunch of nodes allowing you to do traditionally mesecons things directly with digilines instead. 04:40 DeepThgt that was already on the roadmap to add to the server 04:40 DeepThgt just have performance issues to tackle first 04:42 DeepThgt not that i havent optomized all i can engine side, i just have ideas :D I dont want egg on my face so tight lips unless i/we (the other admin) pull it off 04:43 DeepThgt cause if succesful, the backend ideas we have will make a lot of things possible that are currently not due to the architecture of the engine 09:01 MTDiscord The formspec in TypeScript https://github.com/jordan4ibanez/forgotten-lands/blob/main/mods/formspec/init.ts 13:08 erle MisterE123 next blog post when? i want to have xcam in it 13:11 MTDiscord We will probably do one for October-November. It would be nice to mention the game jam in it before it officially starts. 13:12 MTDiscord So... at the end of November. (read: January) 13:18 erle LOL 13:18 erle minetest blog time dilation 13:18 erle thx 13:19 ROllerozxa soon the blog interval is gonna become yearly 8) 13:20 celeron55 it'll take exactly one blog-month 13:38 rubenwardy just rename to "Recently in Minetest" or "Previously in Minetest" 13:38 rubenwardy or "Minetest updates" 13:38 rubenwardy no that's overloaded 13:40 MTDiscord "some time in the past in Minetest" 13:41 MTDiscord or better yet, "Once Upon a Time in Minetest" 13:55 erle or just release very months on a schedule unless you have nothing 13:55 erle small blog posts are nice too 13:55 erle every month i mean 13:55 rubenwardy I think the issue is more available time rather than not having enough to write about 14:02 MTDiscord It's impossible to have nothing because then you'd just have an article about how blog contributions are worse this month than usual and people need to step up their game 😏 14:02 MTDiscord IIRC the blog is designed around a methodology that could scale well, but in practice people submit requests rather than articles so writing and editing bottleneck. 14:06 rubenwardy yeah exactly. Needs more editors 14:07 rubenwardy I'm too busy writing about writing to work on the MT blog 14:07 MTDiscord I'm too busy writing in here about how there aren't enough people writing elsewhere to write elsewhere. 14:12 erle rubenwardy btw would you maybe accept additional info about texture handling? 14:12 erle rubenwardy for your book? 14:12 erle rubenwardy stuff like which other tools exist, how to optimize textures, what texture formats minetest does not load (it does not load all PNG or TGA files, not sure about JPEG, but arithmetic coding sometimes does not work with programs) 14:13 erle my simplest suggestion would be: a bunch of people use mtpaint for textures, but it is not mentioned (the painting program, not the minetest mod) 14:13 erle https://en.wikipedia.org/wiki/MtPaint 14:13 MTDiscord Oh, a "now that you know how to do it, here's how to do it well" section in the book could be nice, with things like best practices and optimization advice 14:13 erle yeah 14:14 erle there is so much shitty optimization advice lol 14:14 erle (mostly in the forums and on bug trackers) 14:14 erle by shitty i do not mean stuff that is outdated, more stuff that totally sounds plausible but does not hold up once you benchmark it 14:14 MTDiscord 🤔 14:16 MTDiscord There are some basics to distribution, like, make a repo, submit to cdb, use .gitattributes to exclude non-functional files from the "to play" audience download, and run your images through a few lossless optimizers, that are kinda obvious. 14:17 MTDiscord One trick I've found is to make your image transparent areas alpha=1 instead of alpha=0, so that optimizers don't throw away the transparent part RGB info. Compare leaves in NodeCore in fancy vs opaque: they look pretty nice in both cases, but I had to preserve RGB info that can't be interpolated to make them work. 14:18 MTDiscord In the long run I guess mod optimization advice could turn into a second book though. 14:18 MTDiscord Maybe Ruben would want to sit that one out for the most part 😂 14:18 erle warr1024 the lossless optimizer stuff is not obvious, which is exactly what you just explain lol 14:20 erle also i have noticed that people are generally not using colormapped images, even though they could 14:21 erle or texture atlases for that matter 14:21 rubenwardy when it comes to optimisation, I think the best approach would be for people to write blog posts or forum topics about different things. Then the modding book could maybe summarise some important ones and cite those posts 14:21 erle rubenwardy and with painting programs? i already know that gimp is not a very popular tool for pixel art or textures in general 14:22 erle (at least for those who actually choose the program instead of using the first thing they have installed, e.g. ms paint or gimp 14:22 erle ) 14:22 MTDiscord Gimp seems to be pretty popular in the sense that a lot of people use it, but not necessarily in the sense that a lot of people want to use it. 14:22 erle better phrased than i did hehe 14:23 MTDiscord I mostly use gimp myself because I've got the head space to fit a jack of all trades tool, not a whole bunch of specialized shit. 14:23 erle last night i told my friend to just check something in gimp and he was like “i am using mtpaint, i do not even have gimp installed now” 14:24 MTDiscord MtPaint is unfortunately named considering our context here, since I assumed it was a mod that let you paint directly on cubes in the game, which is a gimmick I definitely don't need. 14:24 erle Warr1024 those GNU Unifont emojis i did not create with emacs i created using the X11 bitmap(1) program ;) 14:25 erle Warr1024 there is also a mod that recreates ms paint in minetest https://cheapiesystems.com/media/mtpaint-saveload.webm 14:25 MTDiscord Heh, my preferred format for MT schematics is ASCII art in Lua source code, so there's no accounting for taste when it comes to tools and workflow. Gotta have my spacebar heating. 14:26 erle oh, i actually created a bunch of bitmaps for my mods by having lua code generate them with tga_encoder 14:26 erle makes it easier to have a usable diff when i update them 14:26 erle and it's only small grayscale textures as far as i can remember 14:27 erle (okay, almost all minetest textures are small) 14:27 MTDiscord I have a few textures that I "bake" using imagemagick because I use a few effects that texturemods can't do, and so I have some media source that's actually obviously code ... and hence one of the reasons I don't do separate media and code licenses... 14:28 erle what effects are those? 14:29 erle i mean apart from gamma-correct blending :P 14:30 MTDiscord Like blur type stuff I think. 14:31 MTDiscord Also I'm a bit unsure about how heavily I should lean on texturemods at runtime to begin with. I do some stuff with them that's borderline abuse for a base game texture. 14:42 erle Warr1024 please tell 14:43 erle Warr1024 i was thinking about making a pixelops mod that does some blending etc. operations in lua 14:43 erle instead of using texmods 14:45 MTDiscord erle: sounds like https://github.com/appgurueu/modlib/blob/master/minetest/texmod/gen_tex.lua 14:47 erle -- TODO implement counterclockwise rotations to get rid of this hack 14:47 erle for _ = 1, 360 - self.rotation_deg / 90 do 14:47 erle t = t:rotated_90() 14:47 erle end 14:47 erle this is peak lua 14:47 erle luatic when are you going to split up modlib lol 14:47 erle luatic also this is entirely unlike what i want. i want NO texmods. not even texmod parsing, what you seem to do here (right?). 14:48 MTDiscord erle: you can use this without texmod parsing 14:48 MTDiscord just use the dsl 14:48 erle we have now 17 competing stantards lol 14:48 erle gamma-correct blending when luatic 14:49 erle the t/d being directly next to each other on neo2 is the most obvious trap i fall into 14:49 erle repeatelyl 14:49 MTDiscord see https://github.com/appgurueu/modlib_test/blob/master/minetest/texmod/test.lua btw 14:49 erle repeatedly 14:49 MTDiscord oh and also https://github.com/appgurueu/modlib/blob/master/tex.lua 14:52 MTDiscord Basically modlib.minetest.texmod centers around Minetest's texmod abstraction; it allows constructing (using a simple DSL), parsing (efficiently!), writing (efficiently!), calculating dimensions of, and (roughly) computing textures (if base images are provided). 14:53 MTDiscord modlib.tex is just a texture implementation; it is the base modlib.minetest.texmod uses when computing textures. It tries to not have its API influenced by Minetest too much, but because it needs to provide all the stuff texmods need, it can't be too different either. 14:54 erle luatic why exactly do you not make a lot of minimods but stuff everything into modlib? 14:55 erle “(roughly) computing textures” yeah i can immediately see a bunch of cases it does not handle lol 14:55 erle the equivalence between that and the minetest texmod handling is slightly questionable ;) 14:55 erle but i appreciate the effort 14:55 erle have you ever thought about cleaning up minetests texmod handling? you probably did right? 14:56 erle i remember lizzy had some program to get used texmods right? 14:56 erle did you write/use that too? 15:15 sfan5 !seen Wuzzy 15:15 MinetestBot sfan5: wuzzy was last seen at 2023-11-11 14:23:37 UTC on #minetest 15:20 MTDiscord erle: lizzy's program is https://github.com/LizzyFleckenstein03/texmodbot/, yes I used that 15:22 MTDiscord erle: I am aware (1) that it is not fully equivalent to Minetest's parser - it is not intended to be; (2) that computing the textures differs (for example, I didn't bother looking into how exactly Minetest does scaling) 15:23 MTDiscord I have thought about cleaning up Minetest's texmod handling, but C(++) is more work, the entire PR process etc. would take a while, and it would probably also be too much work to convince everyone that it is "equivalent enough" 15:25 MTDiscord as for why modlib is not a bunch of minimods: good question, laziness to some extent I'd say 15:28 erle luatic tbh i would actually use parts of modlib if it was not a big ball of stuff 15:32 MTDiscord " Warr1024 please tell" <-- https://gitlab.com/sztest/nodecore/-/blob/master/mods/nc_sponge/node.lua?ref_type=heads#L40-60 15:33 erle spongebob_face.jpg 15:33 MTDiscord In the case of water animations, I bake them externally from the source material; in the case of the living sponge animation (a subtle pulsing crossfade) I bake the animation via texturemods. 15:33 erle i like that style 15:34 erle much nicer than texmods 15:35 erle where is the source for nodecore.tmod located? 15:42 MTDiscord https://gitlab.com/sztest/nodecore/-/blob/master/mods/nc_api/util_texturemod.lua 15:45 erle Warr1024 nice mini mod 15:45 erle i wonder if it would be nice to have a “mini mod” challenge 15:46 MTDiscord you could, but tbh writing really elegant library code is the kind of sport that appeals to only a very specialized kind of spectator ... and actually adjudicating such a contest sounds like a nightmare 😄 15:47 erle the challenge is to come up with useful code 15:47 erle not to ”win” 15:48 erle i think AOSA has a “X projects in 500 lines or less” booklet 15:48 MTDiscord lines are a useless metric anyways 😏 15:49 erle https://aosabook.org/en/500L/introduction.html judge for yourself 15:49 erle https://aosabook.org/en/ sorry 15:50 erle luatic it is more a ballpark thing. the entire redo build system (well, my implementation) is less than 500 lines. as are a bunch of really interesting minetest mods. tga_encoder is 700 lines, but it contains three separate RLE encoding functions because i was lazy. 15:50 erle one of them is even suboptimal 15:51 erle (some color reduction happens during encoding and is interleaved with RLE, so the reduces color is encoded, but the decision when to encode a new packet is based on the not-reduced color) 18:03 erle sfan5 do i get an explanation? 18:04 sfan5 we haven't discussed what to communicate as the reason/explanation 18:07 muurkha erle: of what? 18:07 erle muurkha this: https://irc.minetest.net/minetest-dev/2023-11-19#i_6134290 18:08 sfan5 I mean there are obviously reasons so I'll ask the rest of the team 18:08 erle i mean, if it is not a permanent ban there is surely something you want me to do 18:09 erle i must say i am not sure what being removed from the org actually implies 18:09 erle i mean, i doubt i ever had special privileges? 18:09 erle i mean, if so, i did not exercize them 18:10 erle (or did not realize i did) 18:11 muurkha oh, condolences 18:32 erle just one more question about it: was *this* internal discussion #71? i always wondered. 18:32 ShadowBot https://github.com/minetest/minetest/issues/71 -- make paths modifiable by hasufell 18:38 MTDiscord heh, thanks bot, for posting external discussion 71 out of context 😄 18:42 erle i just want to point out that if you want me to *stop posting about regressions to the issue tracker* (not sure if that is meant with “minor issues”, but it is a lot of what i do) you can a) ask me to b) ban me entirely 18:42 erle from github, i mean 18:42 erle i mean, i have already figured out that it is not appreciated 18:44 erle (which is why i am generally much more annoying about regressions that affect everyone than stuff that is only happening on my hardware) 18:46 MTDiscord If it's really bothering everyone, then surely somebody else will already be likely to file the issue. Considering that we are not suffering a shortage of issues right now, it's not necessarily a tragedy if one gets missed and we have to discover it again later, either. 18:46 erle ah, now i get it https://github.com/minetest/minetest/pull/14006 18:48 n_to First they came for Erle and I said nothing, because I wasn't Erle. 18:49 ROllerozxa huh 18:50 erle Hey ROllerozxa do you want to adopt that patch lol 18:53 erle ROllerozxa if you do, you have to incorporate https://github.com/minetest/irrlicht/pull/254 (which i hereby offer as a sacrifice) into it 18:58 n_to Musste eben sehr über diese tocotronic Lyrics lachen: 18:58 n_to Wenn die Liebe endet, ist es mitten in der Nacht 18:58 n_to Ein Lichtschein, der mich blendet, dringt aus meinem Tiefkühlfach 18:58 n_to Dort liegt eine Pizza, die ich aufzupeppen versuch 18:58 n_to Mit Kräutern der Provence hab ich keine Chance 18:58 n_to Ah-ah-ah, ah-ah-ah, ah-ah-anse 19:06 MinetestBot 02[git] 04sfan5 -> 03minetest/minetest: Try to fix VS build failures 130e4de28 https://github.com/minetest/minetest/commit/0e4de28988a96e87f6391a654eff8eaeed3ef801 (152023-11-19T15:22:20Z) 19:08 sfan5 erle: I will adopt the PR 19:09 erle ok cool 19:46 MinetestBot 02[git] 04Desour -> 03minetest/minetest: Fix cached wanted_range and camera_fov being written with out-of-rang… 1373e85b2 https://github.com/minetest/minetest/commit/73e85b2ebb3e6649244a6fe4548f5b160c7c207d (152023-11-19T19:45:18Z) 19:46 MinetestBot 02[git] 04Desour -> 03minetest/minetest: Fix undefined inf to s32 cast in GUIScrollBar::setPos 131bc74b0 https://github.com/minetest/minetest/commit/1bc74b0ba1d92c1b19ce281f974d79e549107794 (152023-11-19T19:45:18Z) 19:46 MinetestBot 02[git] 04Desour -> 03minetest/minetest: Clamp values in read_ARGB8 13585e6aa https://github.com/minetest/minetest/commit/585e6aa80ba9ee711aebcf3630ab8bbec1e446fd (152023-11-19T19:45:18Z) 19:46 MinetestBot 02[git] 04Desour -> 03minetest/minetest: Devtest: Fix testnodes bouncy color calculation 137199ee4 https://github.com/minetest/minetest/commit/7199ee4ff89f8990fd6c9eed9499108ccefc887f (152023-11-19T19:45:18Z) 19:46 MinetestBot 02[git] (1 newer commits not shown) 19:46 Desour bim bim bim 19:48 MinetestBot 02[git] 04MisterE123 -> 03minetest/minetest: lua_api.md: Add tick marks to position HUD element 1331ee7af https://github.com/minetest/minetest/commit/31ee7af3ab805b1c7b64b7a03ff2b663b8030265 (152023-11-19T19:46:40Z) 19:48 MinetestBot 02[git] 04Wuzzy2 -> 03minetest/minetest: Fix mod translation updater bugs (#13974) 1361db32b https://github.com/minetest/minetest/commit/61db32beee936a5264ef17785d8efaa7cc14bda4 (152023-11-19T19:46:52Z) 21:03 MTDiscord What does the "Unconfirmed Bug" tag mean? Does that mean that anyone can repro and confirm it by posting a comment, or does it need to be specifically reproable by a Core Dev? 21:06 sfan5 just an indication of state, help is welcome as always 21:07 sfan5 if someone comes around and says they reproduces it (and it's credible), we can re-tag the issue 21:14 celeron55 it could be written out as "failed to reproduce bug or information missing for reproducing" 21:15 fox Hello! 21:16 MTDiscord In this case it sounds a bit more like "I don't have the time to even attempt to reproduce it right now, but am hoping somebody else does." I actually thought we had it as an automatic tag at one point, but it looks now like it's applied manually by triagers. 21:17 sfan5 pretty sure it's still autoamtic 21:17 sfan5 automatic* 21:18 celeron55 it could be renamed if the name is too discouraging or confusing 21:18 celeron55 "Problems reproducing bug" or something 21:19 MTDiscord oh, haha, it IS automatic, I misread my own damn name as wsor4035 because his name was right under it, and our names look too similar. 21:20 MTDiscord It seems okay to me now; knowing that it is actually automatic helps me not take it as some kind of criticism or mistrust or something; it's just a default state. 21:20 celeron55 actually this is not a good name either as it's essentially intended to be applied to all bug issues first 21:21 celeron55 "To Reproduce" would be an usable alternate name 21:21 MTDiscord "Needs verification" or something could be more helpful, i.e. as a call to action. 21:22 MTDiscord some kind of name that encourages people to try it out for themselves. 21:22 celeron55 and of course the reason it's this way and not the other is that a tag gets more attention than no tag, and uncofirmed bugs need more attention than confirmed ones (at least that's my thinking) 21:22 MTDiscord celeron55: when the bugs start to reproduce we'll be in deep trouble... 21:23 MTDiscord I like "needs verification/confirmatioN" 21:23 celeron55 "Needs confirmation" is closest to the current one, people might even realize it's the same tag 21:24 MTDiscord I wonder whether it should be something like "confirmation wanted" rather than saying "need" ... in case we don't necessarily want to imply that it will NOT be worked on unless somebody else can repro it. 21:25 MTDiscord Like, I don't know if anyone will be able to get the same results I did in my tests (people with more powerful hardware will have a harder time of it) but it sounds like the fix for it (batching draw calls) is something that's desired anyway. 21:26 celeron55 the tag/label is referred to by name in the issue template so that needs to be changed also if the name is modified 21:26 MTDiscord 14018 just in case the context is missing. 21:26 MTDiscord I think it's a good thing if our issues sound "needy" :-) 21:26 MTDiscord Quite frankly, if no one can reproduce an issue, chances that it will be worked on are pretty slim... 21:26 MTDiscord Well, what we NEED is fixing of problems, not necessarily putting more prerequisites in front of fixing... 21:28 MTDiscord 14018 actually sounds like the kind of issue that a few people have wanted to fix just out of general principle, so actually knowing it will have a real-world impact and isn't just a theoretical or unmeasurable thing might help. 21:48 sfan5 !mod modlib 21:48 MinetestBot sfan5: Modding Library [modlib] by LMD - https://forum.minetest.net/viewtopic.php?t=22345 - https://github.com/appgurueu/luon 21:48 sfan5 that seems a bit off 21:49 sfan5 !cdb modlib 21:49 sfan5 hmm does MinetestBot not have this 21:53 sfan5 @luatic glad to find out that modlib does not optimize written PNGs better than the engine 21:54 erle well you both did the simplest thing that works 21:54 erle it's not surprising 21:55 erle also optimizing PNGs is not so easy unless you know what is in the texture, you have a combinatorical explosion, with the lowest thing doing it 5 times with different prefilters.