Time Nick Message 05:16 lissobone hare_hare_yukai3: top uno momentos bruh 05:16 lissobone numero uno 06:04 MTDiscord Huh 07:54 kyle56 i have questoin. so when a forum thread is hidden by a moderator, can they still see it and any attachments present on said thread? 07:58 kyle56 because there is a thread was hidden a few months ago. and ever since it was, i wanted to download the attachment that i was too dumb to grab beforehand even tho i kinda figured that due to its controversial nature, it would be deleted. 07:59 erle kyle56 what's the attachment? 07:59 erle if it's the phalluz mod, i have it 07:59 erle i know nothing else controversial 08:00 erle like controversial enough to hide 08:00 erle (maybe i don't know it because it was hidden) 08:00 MTDiscord lol, phalluz... 08:01 kyle56 its probably wishful thinking but i am hopeing that a mod would be nice enough to send me the file if its still on the server. because otherwise its pretty much lost media. and i want to preserve it. i know that it may not be suitable for the forum, but i would like to maybe give it a new home. 08:01 kyle56 the thread was saved in archive.org a while ago. so this is the thread i mean. but the attachment was never saved. https://web.archive.org/web/20190609175714/https://forum.minetest.net/viewtopic.php?f=9&t=14906 08:02 MTDiscord @not luatic might be able to provide you with it when he wakes up 08:03 kyle56 well i could give him my email. tho i dont wanna say it here in this chat. 08:04 erle ROllerozxa the phalluz mode is transphobic because it excludes players named samantha yes i am willing to die on this hill, it needs an update 08:05 kyle56 lol do i even wanna know what that is? the name makes it sound like it has something to do with phallic imagery LMAO 08:06 erle i will not tell you what it does 08:06 erle it is a secret 08:06 kyle56 i would ask for not luatic's email but since my personal email is brand new, i probably can't send messages to anyone without it going to spam. 08:06 erle don't mention it 08:07 kyle56 fair enough lol 08:07 celeron55 kyle56: i can download it and send it to you 08:08 kyle56 celeron55: that would be great thanks! but how do you suggest i receive it? i guess i could put a temporary email alias in this chat. that way i can destroy it if it gets spammed 08:09 erle kyle56 do you know about QUERY 08:09 kyle56 oop. i just got your DM. thank you so much! 08:09 erle kyle56 you can send other people messages on IRC privately 08:09 celeron55 it's a very small mod though, it barely does or contains anything at all 08:10 celeron55 not worth much 08:10 kyle56 i kinda figured which was why i never downloaded it at first. i guess the fact that the thread was hidden made me want it more though. made me curious. 08:11 celeron55 it's a couple of non-interactive nodes and one recipe 08:12 kyle56 well i guess its good to know i was not missing out on much. and that my attempt at a similar mod is much more complete. 08:15 erle wait so this mod does what 08:15 erle devour other players? 08:16 erle that does not sound remotely NSFW 08:16 erle like is there a lot of blood or what 08:16 MTDiscord (do you know what vore is erle) 08:16 erle ROllerozxa are you saying devouring other players is lewd 08:16 kyle56 to some people it is. 08:16 erle ROllerozxa you know you can devour CAKE in some minetest games 08:16 erle it doesn't make you a cakefucker 08:17 kyle56 seems that mod just adds a stomach block and a acid block 08:17 erle celeron55 any idea why this was hidden? 08:17 erle like was it the surrounding text? 08:18 kyle56 thats a good question. all i know is that it was hidden after i necroposted to it showing surprise that i was not the first to use that concept for a mod. 08:21 kyle56 i mean i suppose even if its subject is not inherently sexual, fetishising it may be unwelcome on the forum. which is fine. i have no intention of sharing my mod on the forum or contentdb since i know it would be against the rules. 08:23 kyle56 while it seems that devour thing was never finished, i have put a lot more work into my mod. i wont go into details since its probably not appropriate here. but my mod does allow devouring other players. 08:24 celeron55 the topic was moved to trash 5 months ago. i don't know why. if you double-insist i can ask about it from the moderator in question 08:25 celeron55 probably due to a combination of a necropost and a useless mod 08:27 kyle56 oh yeah i never thought about that. man i really wish more software systems had a delete reason system or something. where it would say like "this ____ was hidden for violating this rule: useless mod and necroposting" or something. i like when there is transparency. dunno if phpBB has that but it would be nice. 08:28 celeron55 phpbb's features haven't been chosen to support actual real world usage. it's just a random bunch of features the developers came up with in a pitch black room separated from the real world 08:29 celeron55 yes, you can supply a reason if you delete a mod 08:29 celeron55 but we generally don't delete mods. we move them to the trash pile 08:29 celeron55 and for moving a topic, you can't supply a reason 08:30 celeron55 eh 08:30 celeron55 s/mod/topic/ 08:31 celeron55 (i should be working so apparently i can't allocate enough brain capacity to this discussion to succesfully come up with words) 08:36 kyle56 well i think i am going to be heading off now. again thx for recovering the file, even if it wasn't really anything worthwhile. if erle does want you to ask the moderator why it was deleted, i'd be interested in knowing too but its not that important to me. maybe i will check the irc logs to see if a reason was ever posted. anyway bye. 08:40 erle tbh if i saw a mod where a user was eating another user i would probably not think it is a fetish thing 08:40 erle but more like AMOGUS 08:40 erle is there an amogus mod btw 10:17 erle MisterE123 do i need to do more than open an issue? like make a PR or so https://github.com/minetest/blog/issues/129 10:19 lissobone i don't feel my limbs 10:20 erle lissobone sudden numbness may be a sign of a stroke. TELL THE PERSON NEAREST TO YOU IMMEDIATELY. show them this link: https://www.cdc.gov/stroke/signs_symptoms.htm 10:21 lissobone nah i slept wrongly 10:21 lissobone my heat is healthy 10:21 lissobone heart 10:21 lissobone just woke up guys 10:21 lissobone hi 10:22 erle oh that is an explanation 10:22 erle i hope it clears up then 10:22 erle still, if you have trouble walking or speaking or become confused, call whatever is the emergency number immediately 10:22 erle though i imagine having trouble speaking would make it a bit weird 10:23 lissobone dw bro i just like to write absolutely random stuff in random chats at arbitrary moments that is nevertheless related to me 10:23 lissobone though i do have trouble speaking nowadays 10:23 erle don't bro me 10:23 erle but i don't worry 10:23 lissobone my brain is worn out by korean 10:23 lissobone y not bro u 10:23 lissobone every human on earth is bro 10:24 erle lol 10:24 ROllerozxa if I get confused while reading code, do I still need to dial 112 10:25 lissobone Fun hacker discovery: our brains kind of work like giant list processors. 10:26 lissobone I found out that people remember things sequentially and associatively: when we try to remember something, we look for links that might lead to that cons cell. There might not be a single link, or we might have a direct one to that very cell and its CAR. 10:26 lissobone In the last case, we remember it immediately. 10:28 lissobone alright, it's 19:30, i slept for like 4 hours and a little more and saw absolute banger dreams, time to work 10:29 lissobone hi 10:31 lissobone i feel slightly drunk on sleeping maybe that's the reason i'm sh***ing in this channel (i'm very sorry if that's actually the case maybe i should cut down on lisp consumption) 10:32 MTDiscord lissobone: no, not quite; yes, our brains work very good associatively (with links), but that doesn't mandate linked lists; graph or tree data structures (which arguably can be represented using lists too) come to us very naturally 10:34 lissobone there ain't no trees or grahps in lisp, buddy 10:34 lissobone only lists, for LISP stands for LISt Processor 10:35 lissobone And thinking in terms of lisp lists, having faith in my brainpower, actually helps remember things faster and think better. 10:35 lissobone hi 10:35 lissobone guyz did u know i'm gonna hack my dad 10:37 ROllerozxa words 10:37 lissobone I don't think I should write here. Bye. 12:08 BuckarooBanzai > BuckarooBanzai where's your minetest python lib 12:08 BuckarooBanzai erle: i don't do python :) i have one in go but it's not really useful except for load-testing a server: https://github.com/minetest-go/minetest_client 12:08 BuckarooBanzai Fleckenstein/Lizzy made one that has more features, just fyi: https://github.com/dragonfireclient/hydra-dragonfire 12:08 erle IIRC the python thing was funny for getting a half-initialized player 12:09 erle that appeared as a flat avatar in mcl2 12:13 BuckarooBanzai uh, are you sure? i can't think how that would work: if the player is visible ingame then all the relevant callbacks should have run already 😕 12:16 erle i'm p sure minetest reordered some login logic some way back because of that 14:09 [MTMatrix] mod for enable creative mode for server admin exists? 14:10 [MTMatrix] also, how I can /give block with predefined param2? 14:23 sfan5 while in item form a node does not have a defined param2 14:38 calcul0n_ the best you can do is use the replacer tool i guess 14:41 [MTMatrix] yes, this workaround best 15:47 erle is there a way to see earlier versions of comments, or at least which user has edited them? the top post here used to say something else: https://git.minetest.land/MineClone2/MineClone2-Plus/issues/21 15:48 erle this is part of what is missing 15:48 erle > To everyone except @ancientmarinerdev -- It's been a pleasure to get to know you, and to work with you. 15:48 erle ancientmariner any idea what that is about? 15:48 rubenwardy yes, click edited 15:49 erle ah thanks 15:49 erle i just want to know because i vaguely remember accusations of users on mesehub editing *other* people's comments and i am not sure if that is even possible. 15:52 erle so i guess kneekoo deleted the inflammatory 2 paragraphs in there (given the vitriol i get why) 15:52 erle then i think the rumors of people editing other people's comments *without making it obvious they did* (like adding “admin comment: this comment was redacted for violating the rules”) may be true 15:55 [MTMatrix] " > To everyone except @ancientmarinerdev -- It's been a pleasure to get to know you, and to work with you." - You know very little about me, if you have a negative opinion of me, I'd have to insist on you questioning where you get your information from. 15:55 erle ancientmariner i do not have a negative opinion based on that comment 15:55 erle but i am pretty sure the person posting has 15:55 [MTMatrix] secondly, stop digging up stuff from months ago that has nothing to do with you. i understand you like to focus on negatives, but i don't care much for it 15:55 erle it just popped up in my inbox 15:55 erle i think because the issue got closed? 15:56 MTDiscord also agree, let old drama stay dead 15:56 erle look i cared more about the gitea edited view than your drama 15:56 erle seriously 15:56 erle i definitely did not *dig it up* 15:56 erle it probably landed in your inbox too 15:56 [MTMatrix] gitea is pretty poor, and the access control is dreadful 15:57 [MTMatrix] i'm aware of what was in that post, and i think anyone involved in it wants it in the past 15:57 erle the only person i'll be wary of based on this thing is kneekoo because he did not give an indication that he deleted the unfriendly parts of the comment (if i understand it correctly) 15:58 erle but let's say i don't like kneekoo anyway because opposed to literally *everyone* else i had a disagreement with, kneekoo would never code something (at least when i had disagreements), whereas every single other person would at least make a proof of concept. 15:58 MTDiscord I think you should give prominent indication of moderator edits if the platform doesn't already. 15:58 erle it does not matter 15:58 [MTMatrix] i talk extensively to kneekoo and we run decisions by each other, i have no doubts over his decision making. on the contrary, i have the utmost respect for the way he conducts himself 15:58 erle luatic i think that too! 15:59 erle ancientmariner yeah but like … can kneekoo actually code? because at some point i suggested that maybe he just can't write lua at all and that's the reason (i.e. kneekoo is not being an asshole, just does not know any better) 15:59 MTDiscord erle: are you being codeist ;) 15:59 erle luatic a bit yes 15:59 [MTMatrix] i don't really take much value in what your opinion of people is, and kneekoo has done more for mcl2 that most people in these last 7 months. he's probably a more stoic and reasonable person myself, and your mud slinging says more about your character than theirs 16:00 MTDiscord ancientmariner: btw, do you know any good tool to profile allocations in Lua? 16:01 erle i take this as: 1. you like kneekoo 2. you don't like me for some reason 3. you don't want to tell me if kneekoo can write lua, so i'll keep the implications to myself 16:01 erle ancientmariner i am sorry for giving the impression i endorse that comment. i just found it weird because obviously you value not alienating community members MUCH more than i do 16:02 [MTMatrix] i respect people based on their conduct, i dislike you for the fact you're negative and you always jump to attacks on people's character for their decisions. technically, you're excellent, it's just a shame you do not focus on how you interact with people 16:02 erle oh, so it's nothing new then 16:02 [MTMatrix] but hey, if you choose to focus and engage positives, it's all water under the bridge. i'd like to like you 16:03 erle ancientmariner want to have a chance to be negative about something? i wrote a unicode text rendering library and look for vile criticism 16:03 [MTMatrix] but i don't think you really care about that, and i think it's best your just focus on your stuff, and i'll focus on mine 16:03 erle ancientmariner it is intended to enable you to have unicode on signs, yes in mcl2 16:04 [MTMatrix] i really liked the look for the signs lettering stuff you did, and it'd probably help the project at some point in time 16:04 erle ancientmariner i'd appreciate your comments on the shape of the API with respect to mcl2 experience https://git.minetest.land/erlehmann/unicode_text 16:05 erle be as negative as you want, the worst response you can get from me is probably “i have researched this a lot for months and i am pretty sure your concern is baseless for $OBSCURE_REASON” 16:05 erle but i would like to actually make API consumers happy, so 16:05 [MTMatrix] also mineclone2 can merge lib-signs from minetest mod? Just this works fine with non-english captions 16:05 erle also i'm pretty sure you'd appreciate having unicode on signs right? 16:05 erle localhost i can do much more than latin 1 16:06 erle i can do astral plane characters and combining characters 16:07 erle localhost look here for example renderings: https://github.com/minetest/blog/issues/129 16:07 [MTMatrix] in all honesty, i'll probably have to review this in the future, i've woken with a migraine and my motivation isn't high right now. i'm gaming and recharging. i have a feeling that my knowledge in this area is limited, so my focus on whether it would be easy to pick up and understand without that level of knowledge 16:07 erle oh it's mostly about the API shape, not a code review 16:08 erle i just want to know if the example code at the top of the readme is okay with you basically 16:08 erle or if there is anything missing that is crucial for your use case 16:08 [MTMatrix] fair enough. i'll bookmark, and try to get to it, but yeah, initial impressions when you showed this a few months ago were very positive 16:08 erle it's basically the same. i had someone tell me last night about otf2bdf 16:09 erle so if i implemented a bdf parser (or someone else did) we could actually render arbitrary stuff on signs lol 16:09 erle from OTF fonts 16:09 erle right now, it's limited to unifont, unifont upper, unifont CSUR and some japanese font i think (well i only load .hex files) 16:09 erle ancientmariner do you think you can give me feedback within the next 6 hours or so if the API looks okay for you? 16:09 erle bc i kinda want to get this on cdb so people start using it 16:11 erle btw, pls don't complain about the io.lines() thing. a bunch of people did and it seems it only is obvious to people actually using the library that not all glyphs come from font files, some might be generated or inline in the source code. 16:13 erle sofar you have signs in inside the box, right? do you have opinions on my unicode_text mod? 16:14 [MTMatrix] documentation looks good, and easy to pick up. i'm not keen on the io.lines tbh but i could wrap that in a function 16:14 [MTMatrix] is there any issues with escaping user input? or would you advocate before sending into your api? 16:14 erle nope, not at all. you just submit UTF-8 encoded text. 16:15 erle i mean so far no one HAS tried to break it 16:15 erle but i am pretty sure the worst you could get was a visual impression of control characters 16:15 erle like, let me see 16:15 erle whatever unifont has at 0x02 if someone sent \2 ig? 16:15 erle appguru / luatic probably knows if this is an issue 16:16 erle so what i do is: 1. i decode the input with the utf8 module to code points. 2. i put the codepoints together into a bitmap. 16:16 erle there is some incomplete bidi magic which you can see in the example rendering 16:17 erle but if i understand it correctly approx. no one is using hebrew or arabic (which renders without connected glyphs) in minetest anyway so far (and it can be improved later) 16:17 [MTMatrix] hopefully the size requirements aren't too onerous, and i'd probably want to make sure unused sign text images are deleted 16:18 erle your signs are probably not bigger than … 64×128 or so? 16:18 erle with most of that being empty space 16:18 erle like you have at most four lines of text and i guess like 16 or so half-width characters per line 16:19 erle with those constraints i can make a very weak guarantee: no sign can exceed the typical size of a map in mcl_maps (they are 128×128 i think) haha 16:20 erle but see the bottom of the readme 16:20 erle for a very simple hack 16:20 erle anyway, i am caring about the API shape right now 16:20 erle performance concerns can be addressed later 16:20 erle make it correct first, then make it fast 16:34 erle ancientmariner i have a simple trick for making sure the unused images are deleted: save the text in the meta of the entity displaying the texture. i do that in xmaps. it means 1. you can delete the textures folder at any time and they can be restored 2. it works in map downloads. 16:34 erle i mean xmaps has to save the actual map 16:34 erle but like a sign entity *probably* can contain its text 16:34 erle then you just have to make sure that you regenerate the sign if the file does not exist 16:35 muurkha I think it's common for people to design APIs that cannot be implemented fast if they aren't concerned about performance at all 16:35 [MTMatrix] good point, and i suppose if changed, delete file generated from it 16:35 muurkha hey, that sounds like a job for a... 16:36 erle muurkha in this one i am pretty sure i am concerned about performance 16:36 erle there is a bit of performance tuning you can still do 16:36 erle the loops are not optimized 16:36 erle but that *can* be done fast 16:36 erle i mean faster 16:39 erle ancientmariner given the amount of people complaining about the music thing, why is mineclone2 music not just a separate modpack? 16:39 erle i mean i am a bit sceptical of the claims that music makes the game not run on their computers 16:39 erle but like, in general, why is it bundled at all if it is not a vital component? 16:41 [MTMatrix] IMHO, it is a vital component. I have heard more positive comments than negative. Even some who were skeptical later praised the quality of it 16:41 erle i see, it's a game design question 16:42 [MTMatrix] I think I heard about 5 negative comments 16:42 erle in my experience, that's massive actually 16:42 [MTMatrix] Mt 5.8 should improve loading of media 16:42 erle like whenever i got 5 negative comments about the samething for software i wrote, i was like “uh what's the issue” 16:42 erle i mean sometimes they are all the same kind of clueless 16:42 [MTMatrix] So its a temporary issue for the handful 16:43 erle but in this case, why do you think that the game runs poorly? (as i said, i am sceptical of the music thing being the reason) 16:43 erle what exactly does mt 5.8 different here? 16:43 [MTMatrix] Many reasons. Poorly optimised. Abms, lbms need investigation 16:43 [MTMatrix] Global steps also 16:44 erle i thought you kinda optimized that? 16:44 erle like i have not tried it 16:44 [MTMatrix] It's a long grind. Made many positive steps. Single player experience is solid 16:44 erle but i remember you being big on monitoring performance at some point in thepast 16:44 [MTMatrix] Network could be giving us prpblems 16:45 [MTMatrix] And my knowledge isn't great on what is causing it. Any pointers in that regard would be really appreciated as I know you've put work into this in the past 16:45 erle having cleaned up network stuff for mcl2 in the past, the only reason for network problems were a) oversized items (which i do not think you can create with new versions of mcl2) b) extremely shitty mods that were developed by 1000 monkeys on 1000 typewriters 16:46 erle ancientmariner i suggest to monitor the raw network traffic while you do something in the game 16:46 erle this is how i did it 16:46 erle like for each packet, find a justification if it needs to be sent 16:47 erle i think the stupidest thing was that some stuff was sent all the time (if the player is sprinting i think) 16:47 celeron55 i would love to watch a video of someone doing that type of optimization work on something running on MT 16:47 [MTMatrix] I tried wireshark and my eyes glazed over. Network is a weakness in my skill set. Wouldn't know where to start 16:48 erle celeron55 it's super fucking boring though 16:48 erle like you have these small things and save a byte here and save a byte there 16:48 erle usually there is no one big thing 16:49 erle this is, by the way, why i focus on small incremental improvements (if the architecture is sound) 16:49 celeron55 well, it wouldn't be for everyone, but it would be much more interesting than a generic programming video 16:49 ROllerozxa "i suggest to monitor the raw network traffic while you do something in the game" how does one do that? 16:49 celeron55 wireshark 16:50 erle ROllerozxa, wireshark and piping the debug output into https://git.minetest.land/Mineclonia/Mineclonia/src/branch/master/tools/analyze-packet-spam if you are lazy 16:50 erle the thing about the small incremental improvements is that sometimes you have a small thing that is *everywhere* 16:51 erle for example, a subset of mod devs seems to think that putting everything in a globalstep makes stuff easier to reason about than having it event-driven (i.e. the code only running when it should) 16:51 erle the problem is, yeah, it is easier to reason about: but it happens in every global step then 16:52 ROllerozxa ah, wireshark. I was wondering if there was anything higher-level, maybe something that would annotate the packets or something 16:52 erle you can get surprisingly far with that shell script i linked (and wrote) 16:52 [MTMatrix] that would be lovely, even on seeing the traffic, it's hard to pinpoint what calls it is and from where 16:52 erle but in the end you may need wireshark or put printf debug statements in the network stack 16:53 [MTMatrix] if there is ways to debug which server -> client calls are made, it could really help 16:53 sfan5 wireshark can annotate them if you have the dissector installed 16:53 sfan5 if you don't need so much detail you can easily just read the debug logs 16:53 erle is the dissector actually maintained? 16:53 erle like it was not always too useful for me 16:54 erle compared to, say, printing from the packet handler 16:54 erle or – sacrilege – using a cheat client 16:54 erle (which is about the same tbh) 16:55 erle i think the funniest thing is that cora found one bug that was actually also in minetest_game (and had been in there forever) 16:55 [MTMatrix] > using a cheat client 16:55 [MTMatrix] but CSM restrictions... 16:55 erle ancientmariner, btw a good first approximation would be to simply measure bandwith 16:55 celeron55 i guess it depends on how low or high level information you need. wireshark is targeted towards low level, MT's debug logs are targeted towards high level 16:56 erle earlier mcl2 versions for example were able to let players be on fire multiple times 16:56 erle and being on fire would create traffic 16:56 erle i think cora managed to have like 400kb/s for one player by being on fire very much but not dying hehe 16:57 erle so ancientmariner i would initially just monitor server-client traffic while playing the game – and at ANY point where you get much more sustained traffic than minetest_game, figure out what could have caused it 16:57 erle with sustained i mean: obv you have traffic when you teleport (nether or end or overworld mapblocks are loading) or when you make a map (the map is transmitted) 16:57 [MTMatrix] i did do a little starting mt in debug mode and managed to find some ely issues, but some of the classifications are quite generic and it's hard to know what is doing them 16:58 erle but when you stand still, for example, you should not have much more traffic than minetest_game 16:58 erle so just start with that 16:58 erle ideally stand still on a server with other people and on a server without them 16:58 [MTMatrix] ah, i guess using mtg as a baseline might be a good idea 16:59 erle yeah, you might have momentarily more traffic than it, but you generally should not have any problem here 16:59 erle for me this was the most common thing that is invisible to server profiling 17:00 erle too much traffic 17:00 erle ancientmariner i also suggest what my ex boyfriend suggested: speedrun the game on the worst machine you can find in your house before every release. a speedrun does not take more than 20 or 30 minutes. 17:01 erle and also you can let someone else do that 17:01 erle you'll notice performance regressions pretty fast 17:01 [MTMatrix] sp is in a good place. we've had good feedback from people on 2 core machines 17:01 erle wdym 17:02 [MTMatrix] singleplayer is smooth, and very playable for the most part 17:02 [MTMatrix] you get a server with 7-10 players, and all hell breaks loose 17:02 erle oh 17:02 erle yeah that had been in the past 17:02 [MTMatrix] mapblocks taking 10+s to load 17:02 [MTMatrix] your database 20GB+? 17:03 [MTMatrix] chest entities not loading in that time also is quite common 17:03 [MTMatrix] sqlite3? 17:03 erle and the steps to solve it are exactly what i outlined: check for traffic spikes, look at debug output (or let the script do it), use wireshark, instrument the network packet handling of the client 17:03 [MTMatrix] these aren't my servers, servers that run the game 17:03 erle localhost it is very unllikely that it is an sqlite problem 17:03 erle most people are not very conservative with their traffic, because they only check singleplayer 17:04 erle i mean i myself introduced a race condition because i only tested in singleplayer 17:04 erle appguru / luatic can confirm that ;) 17:04 erle i need to fix xmaps 17:04 [MTMatrix] singleplayer nick reserved only for local game or possible to use? :D 17:04 erle lol 17:05 erle says the person with the nick localhost 17:05 erle ancientmariner a hint: changing entity attributes might send a network package 17:05 [MTMatrix] yeah 17:05 [MTMatrix] :D 17:05 erle bc of course the client needs to know 17:05 [MTMatrix] yeah, i've been thinking about that when optimising, but couldn't make a dint in the packets 17:06 [MTMatrix] but i think we do need to audit and make it as infrequent as it can be 17:06 erle another thing i can suggest is to disable all mods and enable them one by one. this requires you to have a proper mod structure where mods can be disabled and enabled on their own. 17:07 erle (i.e. the thing i frequently complain about is actually important) 17:07 [MTMatrix] ah, interesting idea 17:08 erle i'm pretty sure if you have not changed the mod structure in the last 6 months or so it is not easily possible 17:08 [MTMatrix] i guess doing so, i'll learn more 17:08 erle like, it's not your fault, almost everyone was dismissing my “if you add something, add it as a clean small mod” 17:09 erle you probably have had to deal a lot with people who prefer big-ball-of-mud mods 17:09 erle because they could not specify a dependency graph if their life depended on it 17:09 erle oh one more thing 17:09 erle hear me out 17:10 [MTMatrix] i think a lot of people don't think about this and why it's important. it's only when you need to do things like this, you realise 17:10 erle ancientmariner i have an idea: laughing obsidian! 17:10 erle :D 17:10 erle yeah, but the problem is that if people are incompetent at systems design they still have opinions 17:10 [MTMatrix] i guess if mojang implement, it'll be another bit of pointless content someone will be desperate to add in 17:10 erle hehehe 17:11 [MTMatrix] i don't like to blame folk. i don't think you realise how rich your experience is. it isn't other people's fault they aren't at your level 17:11 erle nah 17:11 erle but arrogance is a fault 17:11 [MTMatrix] it's just finding ways that people can develop that knowledge 17:11 erle (look who is speaking, me, i'd never be arrogant hahahaha) 17:12 [MTMatrix] i have joked that i used to think open source was built on altruism. i later learnt it was built on ego 17:12 erle the easiest way that i have seen people develop knowledge is to reject their patches with appropriate comments on why they are bad, so they try again next time. i know you do not like this very much, as it is – i must admit – a bit mean. but it is knowledge transfer. 17:12 erle it can also be demotivating 17:12 erle but sometimes i think that is for the best 17:13 erle in early minetest someone implemented a whole weather system 17:13 erle but that person did not adjust their code correctly 17:13 erle so obv it was more of a drive-by PR 17:13 [MTMatrix] i don't object to it, and i tend to block stuff that takes things backwards. i review harder for people that have had a few prs in, so build people up one step at a time 17:13 erle shoot in the general direction of the project and press the trigger 17:13 erle then drive away 17:14 [MTMatrix] main reason i jump in now is finding nodes or objects near on global steps 17:14 [MTMatrix] it's horrific on the cpu and done far too frequently 17:14 erle do you know about semgrep 17:14 erle you can code patterns to find this kind of thing automatically 17:14 erle i used semgrep to find dupes 17:14 [MTMatrix] i agree with your assessment that things should be event driven where possible 17:15 [MTMatrix] i'm not familiar with semgrep, but then i've also never tried luacheck (i did try to set it up, failed badly and didn't care enough) 17:15 MTDiscord wait, MCL is not using luacheck? 17:15 MTDiscord MCl2* 17:15 [MTMatrix] some do 17:16 [MTMatrix] afcms is big on it, for example 17:16 MTDiscord Not mandating static linting when the codebase is this large and you have many contributors is a pretty bad idea tbh 17:17 [MTMatrix] but then again, static code analysis still needs brains applied to it. i've seen prod bugs introduced at work because someone fixed static code analysis bugs and broke crucial functionality 17:17 erle ancientmariner regarding these people trying to optimize the music, i strongly suggest you block every single one of those PRs unless they can explain what the nyquist frequency is and what commands they used 17:17 MTDiscord ancientmariner: of course, but having it in the first step is crucial 17:17 MTDiscord lua allows you to make so many dumb mistakes that luacheck will happily catch 17:18 erle ancientmariner what you do with semgrep is not like a normal linter. you fix a bug and then code a pattern for that problem to never introduce it again 17:18 erle ancientmariner: like here where i abstracted the code pattern that enabled ghost dupes https://git.minetest.land/Mineclonia/Mineclonia/issues/292 17:19 [MTMatrix] i agree it's useful, and probably something we need to do more of 17:19 [MTMatrix] i have a feeling if luacheck was run on mcl2, computers would explode ;) 17:19 erle as you can see my pattern finds every block that can be used to ghost dupe 17:19 erle even though the code is kinda different 17:20 erle and there are no false positives either 17:20 erle you should do this for every bug that has a “shape” 17:20 erle i.e. also for every performance fix 17:20 erle “you” as in “every single one of you“ 17:21 erle luatic what's newton's first law? 17:22 MTDiscord was it actio = reactio or conservation of momentum 17:22 erle no credit for partial answers :P 17:22 erle luatic watch this: https://onion.tube/watch?v=OPRIUJzmkC0 17:22 erle it will amuse you 17:23 MTDiscord impressed....maggots 17:24 erle luatic there is also this amazing rap song about kalman filters: https://onion.tube/watch?v=_LjN3UclYzU 17:25 erle the flow is unreal 18:02 mazes_83 wow I wouldn't think there was songs about kalman filters 18:06 muurkha it doesn't mention Kalman filters at all, even a little bit 18:06 erle muurkha wait what is the song about then 18:08 erle muurkha i always thought what it was describing was a kalman filter 18:08 erle muurkha, what is the actual name of the thing it describes 18:09 muurkha maybe negative feedback? 18:09 muurkha the text is mostly a joke, though 18:09 Krock it is somewhat about Kalman filters 18:09 Krock because that's a method to make an educated guess about an object's position or speed 18:10 erle the text is very repetitve 18:10 muurkha it doesn't mention noisy measurements, Gaussian distributions, prior and posterior distributions 18:10 erle it is from a 1997 air force text about missile guidance i think 18:10 muurkha the text applies equally well to navigation using dead reckoning, for example 18:10 muurkha or to navigation using particle filters 18:11 erle well i see 18:11 erle it's not specific enough 18:11 muurkha right, it describes a problem to which a Kalman filter could be applied 18:11 muurkha but it doesn't describe actually applying the Kalman filter to it 18:11 muurkha or even the conditions that would make a Kalman filter a particularly good solution 18:11 rubenwardy a kalman filter does let you know where the missile is not 18:12 rubenwardy assuming that's what the video is about 18:12 muurkha only in the sense that it lets you know where it is 18:12 jonadab Someone should write a song about the non-trivial zeroes of the zeta function. 18:12 erle the text is from here, on page 5, unter the heading “GLCM GUIDANCE SYSTEM” https://web.archive.org/web/20160903175603/http://afmissileers.com/newsletters/NL1997/Dec97.pdf 18:12 muurkha https://knowyourmeme.com/memes/the-missile-knows-where-it-is 18:14 muurkha as explained above the text, "it may not be the first time you have seen this — [it] seems to apply to all guidance systems" 18:14 erle i see that you are right 18:16 erle jonadab the green party in germany traditionally hosts an “ask us anything” before elections and multiple times i have seen the question “do all non-trivial zeroes of the riemann zeta function have the real part ½ ?” 18:17 erle the song is actually about how we are nothing to the missile but just another target 18:18 muurkha that's the spirit 18:19 erle the missile has studied gorilla warfare and graduated as one of the top of its class or something 18:20 * muurkha studies chimpanzee warfare 18:22 erle i am not credible enough to join the 3000 femboy propaganda officers of NAFO ._. 18:24 erle celeron55 the problem with extremely boring coding stuff is i think also that it takes a long while 18:24 erle celeron55 like lizzy & me took like 9 hours to figure out the stencil shadows thing completely 18:24 erle or something like that 18:25 muurkha someone who doesn't know you might think 'femboy' is pejorative 18:26 erle you are the second person claiming that 18:26 erle to me 18:26 erle like in which part of the world is femboy not an exaggerated opposite of tomboy? 18:29 erle old but good “you are likely to be eaten by a grue” https://onion.tube/watch?v=4nigRT2KmCE 18:29 muurkha people often think 'tomboy' is pejorative, but because patriarchy, only mildly. hierarchical structured oppositions blah blah blah 18:32 erle muurkha, the left must be forcefemmed! https://content.queer.party/media/cache/media_attachments/files/111/017/534/605/268/832/original/1ddbad52611c58c5.png 18:34 muurkha hahaha 18:35 muurkha anyway you can get beaten up for being a tomboy but not usually raped. being a femboy often gets you robbed, raped, or straight up killed. depending on social milieu, of course 18:36 erle i live in berlin though 18:37 erle wearing a black miniskirt gets some positive attention from women between 20 and 30 and rare negative attention from really old men 18:40 muurkha yeah, a thing I really appreciated about San Francisco was that I no longer had to worry about getting beaten up for being gay 18:41 muurkha the funny thing about that is that I've never so much as had a dick in my mouth. nobody cared 18:42 erle i had this funny experience in cologne earlier this year, i asked a woman for the way to the subway station 18:42 erle and she told me the way and told me that she totally knows that i'm definitely not straight 18:42 erle how do you know i asked her 18:42 erle “i'm a lesbian, i can tell!” 18:42 erle weird, but amused me until i fell asleep that day 18:43 muurkha heh 18:50 erle the weirdest case of negative attention i got was when me and my then-bf were called a homophobic slur in amsterdam … in a gay street next to some gay establishment. like, yeah, uh. who else would hang out there? 18:51 erle it was obviously not dangerous, but just super weird 18:54 sofar erle: no comment on the unicode_text thing. But I don't think I understand the mechanics fully. Do the bitmaps get generated offline? Does the server need to be restarted after bitmaps are generated? 18:55 erle sofar it is simply a pure-lua unicode rendering library. unicode text goes in, bitmap falls out. 18:55 erle i mean, it's actually more nuanced than that 18:55 erle but that's the gist 18:55 sofar so you'd have to code a different mod to attach these textures to nodes etc. okay. 18:56 erle yes, this is purely an expression of my disgust with the way people piece together signs out of dozens of small image files holding a single character each 18:56 erle which obviously does not scale as well as … uh … loading a font 18:56 erle (to all unicode characters, i mean) 18:56 erle sofar the end goal is to have unicode on signs 18:57 sofar one of the ways to improve things is to make media content dynamically loadable, barring that you're achieved it I think 18:57 erle out of scope 18:57 rubenwardy we've had dynamic media for a few releases 18:57 erle i just render to texture, dynamic media exists 18:57 rubenwardy just not on nodes 18:58 sofar entities? 18:58 erle sofar i have written an extremely shitty unreleased mod that uses it to make textures for a bunch of nodes to show off that i can do the full range of unicode (well, whatever is in the supplied font) https://github.com/minetest/blog/issues/129#issuecomment-1705650724 18:58 erle cora has already made a thing that puts it on entities 18:58 sofar I read the github issue 18:58 erle also, signs actually often have entities on them for the text 18:58 erle this is funny because you can use pistons sometimes (depending on the game) to push the text away from the sign 18:59 erle so it floats in thin air 18:59 sofar signs using entities are cringe 18:59 erle i used to do that as part of my mcl2 “make the servers weird” gameplay 18:59 erle sofar how do you do it non-cringe 18:59 sofar fix the engine to support dynamic server side textures 19:00 erle yeah, but *what* would have the dynamic texture 19:00 erle something must have it 19:00 sofar rubenwardy: so could you currently dynamically modify/load/update player model textures? 19:00 rubenwardy yes 19:00 rubenwardy there's an API to send new media files to clients at run time 19:00 rubenwardy players and entities have dynamic definitions, so can use dynamic media 19:00 sofar Cool, I have to look at that then 19:00 rubenwardy nodes have fixed definitions, so the texture name cannot change 19:01 rubenwardy you cannot update or remove media currently, only send new media 19:01 erle sofar it's a bit hard to get right for all cases because the api has 3 different versions. ask appguru for details or look at this thing ig: https://git.minetest.land/erlehmann/xmaps/pulls/2 19:02 sofar rubenwardy: seems like a sane trade off. 19:03 erle sofar basically, the same function for sending media exists in minetest 5.3, 5.4, 5.5, but: it takes one parameter in 5.3, two parameters in the 5.4 and suddenly became asynchronous in the last. you can probably see how this is … not the best API upgrade design (it lead to a lot of race conditions). in any case, you can *probably* just ask appguru / luatic if you can reuse that code from the PR for xmaps (that i still need to tes 19:03 erle t and hopefully merge). 19:04 erle sofar the most common error (i made it twice even) seems to be to assume that if one player received media, all received it. at least that is what i take from the PR. 19:06 erle sofar, IIRC appguru / luatic also had this idea to just kick all players that have pre-5.5 minetest to use this feature, which is easier to program and safer, but i consider it a dick move 19:07 erle sofar as far as i can see the most widespread use of dynamic media is mcl_maps from mineclone2 and mineclonia. disregard the mineclone5 version … 19:07 MTDiscord I think sofar presented an interesting idea which may yield a better solution so long as no fancy effects that texmods can't achieve are desired 19:08 MTDiscord you could send glyph textures on demand, then piece together texture modifiers based on those; glyphs that have already been sent wouldn't need to be resent 19:09 MTDiscord so for example if "hello world" is to be generated, then a glyph texture containing the unique letters "eolwrdh" will be sent to everyone, and the server would memorize that it has sent these glyphs and under which texture name and at which positions they are 19:12 MTDiscord sounds like bitmap font snippets 19:13 MTDiscord yeah 19:15 MTDiscord there are various variations of this caching thinkable, such as dividing the code point table into segments of a fixed length (say, 256 glyphs), and hoping that there is enough locality - not sure how well that would work, but it would allow the dynamic media to be cached on disk (since these "pages" of glyphs won't vary) 19:16 sofar For signs I'm envisioning just sending Fonts and letting client side render signs on nodes (instead of entities). That's what I had in mind years ago when I dreamed up a better solution. 19:27 erle sofar i have lookad at all of these and determined that the simplest (and most backwards-compatible) solution is rendering a unicode to a texture 19:27 erle i mean i have looked at all proposals that were issues 19:27 erle and yes, i thought about the send-a-font-as-a-giant-texture thing 19:28 sofar Just send the TTF file or whatever fonts are in 19:28 erle .hex lol 19:29 erle i think i need a FAQ entry for “but the engine should handle it”, maybe it should, but it does not and it can not *possibly* achieve the “edit the pixels table after you rendered the text” thing. 19:29 erle like i can do text effects (you see the color thing) on the cheap. like embossing or whatever 19:30 erle i believe you can send 2 giant pictures that are maybe 2MB each (if you are lucky) for unifont, but *good luck* lol 19:31 erle sofar do you have any comments about the actual API i have? 19:33 erle luatic i have no re-read your (?) proposal and i am really astonished at the level of architecture astronautics hehe 19:34 erle > for example if "hello world" is to be generated, then a glyph texture containing the unique letters "eolwrdh" will be sent to everyone, and the server would memorize that it has sent these glyphs and under which texture name and at which positions they are 19:34 erle this is comedy gold 19:35 MTDiscord I was just spitballing ok :P 19:35 erle nah i love it 19:35 erle it leads to players creating adversarial signs 19:35 erle hehe 19:35 [MTMatrix] who make mapserver for minetest exists in this chat? 19:37 erle localhost, if you want an ingame map, use xmaps or mcl_maps (i am the author of xmaps). if you want an out-of-game online map for a browser, use https://github.com/minetest-mapserver/mapserver and ask … BuckarooBanzai ? 19:41 [MTMatrix] I just want ask how make hard limited map border (no scroll outside this border)... 19:46 [MTMatrix] Or how make this limit in minetest world (no walk/fly/etc outside), no make walls, just really hard limited world include air zone 19:47 erle localhost there is some map border mod on cdb. it will not do *exactly* what you want, but i think it can come pretty close 19:48 erle localhost try it 19:58 [MTMatrix] wait a sec 19:59 [MTMatrix] MapFactory.js; i find it here, just maxBounds 19:59 [MTMatrix] this not ideal, but works 20:35 jonadab A perverse part of my brain kind of wants to do the opposite (make the coordinates 64-bit), despite how impractical such a large world would be. 20:38 erle jonadab the near-term impractical thing is that you need to make something else 20:39 erle already positional calculations near the map-border are not too accurate 20:39 erle even though lua is using doubles internally i think 20:40 erle unfortunately, a proposed refactor towards the goal of bigger maps was summarily executed some time back. did someone necro it since then? 20:40 jonadab Not to my knowledge. 20:40 jonadab In practical terms, it isn't necessary. The existing world size is adequate even for quite active public servers. 20:41 MTDiscord erle: I think the problem is that the engine uses floats, not Lua using doubles. 20:41 jonadab But, you know, that doesn't stop my geek brain from thinking "it would be so cool if..." 20:42 erle luatic the wasn't the refactor about it 20:42 erle jonadab, look at the 3 PRs that tried just that 20:42 erle making the world bigger is a thing 20:42 erle but you need to fix a lot of stupid coordinate things 20:42 erle and also make sure that the behaviour at the map border stays the same 20:42 jonadab What would be more practical, and probably not very much more difficult, would be redesigning the game so that nodes are a cubic foot instead of a cubic yard, or put another way, so that the player is 5-6 blocks tall instead of 2-3 blocks tall. This would completely eliminate the need for stairs and slabs. 20:43 erle LOL 20:43 erle > a cubic foot 20:43 erle what brainrot unit is that 20:43 erle i mean it is a bit funny 20:43 erle because the player avaters probably have … cubic … foots 20:43 erle jonadab, i think you can already grow and shrink players, there are mods for that 20:45 jonadab A number of existing modules would need to be adjusted, including default. (For example, tree trunks would probably end up being made of wood blocks surrounded by bark blocks...) 20:46 jonadab But at that scale, it would be reasonable to let players step onto a block without jumping, meaning, stairs and slabs would not be needed at all. 20:46 jonadab Which would cut a LOT of fat out of the block list in the inventory screen, with no loss of functionality. 20:47 jonadab Very large structures would take longer to build, though. 20:47 erle jonadab, make a game like that go go go 20:48 jonadab Don't tempt me. 20:48 MTDiscord erle: the refactor was about allowing switching from f32 + s32 to f64 + s64, using typedefs i think 20:49 erle luatic and it was killed death