Time Nick Message 00:12 Calinou_ Krock: I created the account requested, just to clarify :) 02:17 VanessaE TIL: Dreambuilder has over 23000 nodes now. wow. 02:17 VanessaE (much nodes. very RAM. wow.) 02:22 MTDiscord https://cdn.discordapp.com/attachments/749727888659447960/826642540891340870/unknown.png 02:30 rubenwardy https://twitter.com/rubenwardy/status/1218257313003458561?s=19 11:17 MinetestBot 02[git] 04sfan5 -> 03minetest/minetest: Irrlicht support code maintenance 131e4913c https://git.io/JYgeb (152021-03-31T11:15:47Z) 13:54 mohanwavale Hello, how to generate dungeon in mod? Where i can read about it? 14:04 MTDiscord dungeons are made by mapgen typically 14:04 MTDiscord i suppose you could make a lua implementation 14:04 MTDiscord https://github.com/minetest/minetest/blob/master/src/mapgen/dungeongen.cpp 14:07 mohanwavale Is it right to do it with abm? 14:09 MTDiscord well it depends if you want to use mapgen dungeons, make your own lua mapgen, or make a scheme adder based on some params 14:11 MTDiscord mohanwavale: usually you don't want ABMs for this, you rather want on_generated 14:11 mohanwavale Thanks, where i can read more about modding then official modding book? 14:14 rubenwardy lua_api.txt is the reference 14:16 mohanwavale Thank you 14:21 celeron55 if you don't like browsing a text file, this is pretty good https://minetest.gitlab.io/minetest/ 14:22 celeron55 (a rendition of lua_api.txt into a reasonable browser format) 14:26 mohanwavale Thanks 14:30 mohanwavale Is there any API mod for mapgen structures? 15:20 tech_exorcist where can I ask a question about a thing that could be abused by cheaters? 15:21 tech_exorcist a thing in the client's code 15:21 tech_exorcist *about 15:22 tech_exorcist not like I want to use it myself eh :-) 15:40 sfan5 here 15:50 tech_exorcist ok 15:51 tech_exorcist apparently, in src/client/client.h, if you modify something::checkPrivilege to always return true, you can use the fly, fast and noclip modes without the necessary privileges, along with the wireframe feature, on survival servers 15:51 tech_exorcist it's just 1 line of code to change, and then several minutes of recompilation time 15:51 MTDiscord yup 15:52 MTDiscord altho about the only thing minetest anticheat is good at detecting is fast, so it will reset you if you try and use it 15:52 tech_exorcist I noticed 15:52 tech_exorcist but it only works horizontally, iiuc 15:53 tech_exorcist and, with several other modifications, it is possible to have everything (except particles, apparently) be lightened by the maximum possible amount of light 15:53 MTDiscord yup 15:53 MTDiscord see df 15:53 tech_exorcist "df"? 15:53 MTDiscord dragonfire 15:54 tech_exorcist ok 16:09 MTDiscord Everything rendering-wise is possible 16:22 eugenefil hi, /rollback command has 2 syntaxes: " []" and ": []" 16:22 eugenefil first syntax is to revert player's actions 16:22 eugenefil what about the 2nd syntax and the actor? can actor be something other than player name? 16:22 eugenefil i looked at engine code and i saw the actor set with RollbackScopeActor and it was used only with player names, i.e. only player can be an actor. am i missing something? 16:44 sfan5 it's entirely possible that what you're thinking of was never implemented 16:46 eugenefil like reverting entities' actions? 16:49 eugenefil i got the impression, that only player's direct actions (e.g. digging a node, taking smth out of a chest) are registered 16:50 eugenefil but launching a grenade or detonating a tnt is non-reversible 16:52 eugenefil i mean the destruction made by a grenade or tnt 18:34 DivideByZero I'm trying to build the latest git version of Minetest client, and I'm havig an issue with the irrlicht. Steps taken: remove system irrlicht, git clone minetest fork of irrlicht, make and build, then git clone Minetest and make and build. If gives an error about finding the wrong irlicht (even though the system irrlicht was removed), and doesn't find the correct irrlicht. 18:34 sfan5 did you tell it where your local copy of irrlicht is? 18:35 DivideByZero No, where is that done? 18:35 DivideByZero from minetest folder it is ../irrlicht 18:35 DivideByZero Not sure what file to place that info in though. thanks. 18:35 sfan5 invoke cmake with -DIRRLICHT_LIBRARY=$PWD/../irrlicht/lib/Linux/libIrrlichtMt.so -DIRRLICHT_INCLUDE_DIR=$PWD/../irrlicht/include 18:36 sfan5 (yes that should be in the README) 18:36 DivideByZero Ah thank you. 18:38 VanessaE if MT requires its own version of Irrlicht now, why is it not simply included within, and built along with MT the usual way? 18:39 rubenwardy Fetchcontent or submodules would be better 18:39 VanessaE submodules suck. 18:39 sfan5 submodules suck and build systems downloading dependencies themselves (by default, that is) is the worst misfeature ever invented 18:39 rubenwardy Including a massive amount of third party code is messy and makes backporting harder 18:40 rubenwardy Most modern languages have integrated tools for managing dependencies 18:40 VanessaE rubenwardy: which begs the question - what's so critical that MT just *has* to have its own Irrlicht fork now, when it's been fine up till now? 18:40 rubenwardy Rust, for example, allows you to give a git URL 18:41 rubenwardy VanessaE: it's not been fine. Input is broken across many platforms, and there are so many graphics bugs that are unfixable 18:41 VanessaE unfixaable when using upstream I guess you mean 18:41 rubenwardy macOS is completely broken. Linux clipboard is broken. Non qwerty doesn't work 18:41 sfan5 was* 18:41 sfan5 (clipboard is still broken but I know the fix) 18:42 rubenwardy I mean with upstream irrlicht 18:42 sfan5 yes 18:42 sfan5 you're correct then, I mean 18:42 VanessaE well fine, if that's how it has to be but DBZ's issue isn't the first time it's come up 18:43 sfan5 that's because we're lacking documentation 18:43 VanessaE n o\ 18:43 VanessaE no. 18:43 VanessaE it's because you're packaging it wrong 18:43 VanessaE to put it bluntly. 18:43 sfan5 the average does not know how to discover existing cmake options from the CLI and the error message or doesn't tell them. 18:43 rubenwardy With a unique name, you can globally install our irrlicht and not need to specify the location 18:43 sfan5 how is that not a documentation problem? 18:43 VanessaE it's like splitting out a function to a separate file and then...adding nothing else. 18:43 sfan5 average user* 18:44 sfan5 or README* (shoulda proofread that) 18:44 VanessaE this MT irrlicht should be packaged with Minetest and should be built with one command (in essence), the same way MT already ios 18:44 VanessaE is* 18:44 VanessaE that's how it's not a documentation issue 18:46 VanessaE sorry if I'm a little coarse.. 18:46 MTDiscord I second the VanessaE opinion here, already talked about that and managed to build statically irrlitch to build minetest 18:47 VanessaE the user should not HAVE to "discover" new cmake variables at all 18:47 MTDiscord Unless in the short term you guys plan to get other people building on top of your Irrlitcth fork, incorporating as part of the codebase makes more sense to me. 18:47 VanessaE it should just default to the copy that ought be included with the engine (note I say "with", not "in") 18:47 MTDiscord Am I the only one who didn't have his build process broken? 18:47 MTDiscord Not sure. 18:48 MTDiscord VanessaE: Irrlicht is even heavier than Dreambuilder 18:48 sfan5 including it inside the engine repo is out of question, everything else can be discussed 18:48 MTDiscord I'm working in Debian and it wasn't easy to make it work, the issue is not only building, also delivering 18:48 MTDiscord We don't want to include that inside the engine repo 18:48 sfan5 what would be possible e.g. is to have minetest detect if you have cloned irrlicht to ./minetest/irrlicht/ and automatically import that 18:48 VanessaE appguru: so? it still has to be built whether included with the engine or not 18:49 sfan5 @Ronoaldo I expect that at least some distributions will make a separate package out of irrlicht-mt 18:49 MTDiscord sfan5: having that makes a lot of sense to me, sort of how it is documented that you can clone minetest_game too 18:50 MTDiscord I sincerely doubt Debian will... 18:50 VanessaE also DB is 168MB.. oof 18:50 rubenwardy That's essentially manual submodules, and also allows systemwide installs 18:50 MTDiscord oof 18:50 MTDiscord that means Irrlicht is most likely smaller than DB 18:50 MTDiscord I strongly suggest static linking... 18:51 DivideByZero I'm getting other build errors now: 18:51 DivideByZero [ 24%] Built target GenerateVersion 18:51 sfan5 @Ronoaldo why not? I haven't read them but I'm pretty sure Debian policies mandate shared linking whereever possible 18:51 DivideByZero make[2]: *** No rule to make target '/home/rick/irrlicht/lib/Linux/libIrrlichtMt.so', needed by 'bin/minetest'. Stop. 18:51 DivideByZero make[1]: *** [CMakeFiles/Makefile2:454: src/CMakeFiles/minetest.dir/all] Error 2 18:51 sfan5 did you check that the file actually exists? I just made an educated guess 18:51 MTDiscord It would create two separate shared libraries that have a lot in common, given irrlitch upstream is already packaged 18:51 VanessaE I'm sorry, whoever made the decision to fork irrlicht and cause users these ^ kinds of issues instead of just including it with the engine made a DUMB decision. 18:52 sfan5 I don't think there's anything else using Irrlicht so distros could just drop that ;) 18:52 VanessaE wrong. 18:52 VanessaE I've run into a couple of other games that use it 18:52 DivideByZero Thanks sfan5. I see the issue: libIrrlichtMt.so is now libIrrlichtMt.a 18:53 VanessaE I keep thinking Supertuxkart is one, not that anyone plays that :P 18:53 sfan5 whether it's .so or .a depends on your settings 18:53 DivideByZero I'll re-make and try again. 18:53 sfan5 VanessaE: supertuxkart has forked irrlicht a long time ago, it's not even named the same anymore 18:53 VanessaE my point stands. 18:53 rubenwardy Does the findirrlicht support IRRLICHT_ROOT? That's usually the more automatic way to find stuff 18:53 sfan5 you mean IRRLICHT_SOURCE_DIR? not anymore and that didn't really work before either 18:54 rubenwardy The old way, though. The new way is to not have find files, and use IRRLICHTDIR to the Config 18:54 sfan5 hm, ubuntu doesn't seem to have a "who uses this package" list 18:54 rubenwardy LIB_ROOT was fairly standard naming 18:54 sfan5 on Arch Linux however there is exactly one (1) package that depends on it: minetest 18:54 DivideByZero changing to .so to .a seems to have worked. 18:57 MTDiscord You can apt or dpkg rdepend, or just look on online packages.ubuntu.com 18:57 sfan5 I looked there: https://packages.ubuntu.com/groovy/libirrlicht1.8 18:57 sfan5 don't see any reverse-deps 18:58 MTDiscord Ubuntu is weird, there were a page in Debian for this. 18:58 rubenwardy Cloning into the root folder, and also supporting a different target name installed system wide makes sense to me as a solution 18:58 MTDiscord Let me check 18:59 rubenwardy The error message can direct users to git clone 19:09 MTDiscord sfan5: it could be my memory failing about the web version. I did on my debian testing install a recursive reverse depends and indeed only minetest seems to use it, along with another tool I guess: https://paste.debian.net/1191851/ 19:09 sfan5 I see 19:09 sfan5 interesting 19:11 MTDiscord That said, for debian, they would usually go either way: dropping the cegui-mk2-0.8.7 support for the next/next release or creating an "alternative" system where you can install one of the forks at a given time... less likelly they will allow both to be installed with different names but that is also a possibility. 19:12 MTDiscord What I'm saying is based on a few discussions on mailing lists and package bug reports I saw, plus my reading of the policy as well as how I see how they support other packages in a similar fashion (like ffmpeg and libav, or imagemagick implementations) 19:13 sfan5 irrlicht and irrlichtmt don't claim to expose a compatible API so I don't think that applies 19:13 MTDiscord Also, this particular library looks like exactly what you guys forked irrlitch (based on the package description from Debian): https://www.cegui.org.uk/ 19:13 MTDiscord > irrlicht and irrlichtmt don't claim to expose a compatible API so I don't think that applies 19:13 MTDiscord fair enoght 19:14 MTDiscord but security issues derived from both do... this is usually where the policy hits harder depending on how much work the maintainers expect to have to maintain both packages 19:15 MTDiscord Perhaps renaming the library everywhere possible? (if it's not done, I may have missed that) 19:16 sfan5 hm 19:17 MTDiscord Food for tough: minetest claims itself to be a voxel game engine, and the fork puts you guys more in that direction. I don't see with bad eyes but incorporating and fiction and improving the internals is good; kindly consider making users lives easier so we all can contribute to that goal 19:25 rubenwardy Irrlicht is a rendering library. Having it included or separate doesn't make a difference 19:25 rubenwardy In terms of definitions 19:27 rubenwardy The problem with including it directly is that is makes the repo much heavier with code we didn't write, makes it harder to backport changes, loses history, and makes it harder to go back to upstream irrlicht 19:27 rubenwardy The majority of users use binaries of Minetest. It can be made easier to build without being directly included 19:28 sfan5 the users who don't use binaries are the most present on IRC though so if process doesn't improve we'll hear many more complaints/support requests 19:34 MTDiscord You don't have to lose history to merge it into MT. But making it a submodule or something would probably make it a lot easier to build, even if you want to keep them separate. 19:40 VanessaE [03-31 15:27] The problem with including it directly... 19:40 VanessaE If upstream is as broken as you say it is, do you really think they're gonna accept changes from MT's fork (or anyone else)? 19:40 VanessaE s/you/you guys/ 19:41 VanessaE as for the size of the repo, no one cares if it's 50 or 500 MB. 19:41 VanessaE it's not like users have to re-clone the entire thing every time they update, if they're not stupid anyway :P 19:47 freshreplicant[m I just noticed - the minetest_game Github page still says: " Minetest Game - The default game for the Minetest engine". I suppose it still is right now> 19:48 freshreplicant[m * I just noticed - the minetest_game Github page still says: " Minetest Game - The default game for the Minetest engine". I suppose it still is right now? 19:49 sfan5 yes 19:51 VanessaE Dreambuilder should be everyone's go-to ;) 19:54 freshreplicant[m Out of curiosity, is there any chance at all the feature freeze for MTG will be lifted at some point, or is it permanent? 19:58 MTDiscord If someone wants MTG to be actively developed again then they should fork it and actively develop it. 20:01 Noisytoot freshreplicant[m: your name contains an unmatched bracket 20:02 MTDiscord https://xkcd.com/859/ 20:02 MTDiscord > VanessaE: supertuxkart has forked irrlicht a long time ago, it's not even named the same anymore sfan5: and they have bundled the lib inside their repo ? 20:02 freshreplicant[m If someone wants MTG "> Makes sense I suppose. 20:02 katp32 hi, this is a bit of an odd question, but is it possible to see somewhere how long a world has been running for? 20:02 MTDiscord https://github.com/supertuxkart/stk-code/tree/master/lib 20:02 freshreplicant[m My name has no brackets in it, to my knowledge, could be some odd Matrix > IRC thing. 20:03 katp32 like, total playtime for a world 20:03 MTDiscord In effect MTG already HAS been forked and actively developed, multiple times. A lot of the games that are out there now for MT are built on top of an MTG base. 20:03 MTDiscord Matrix adds [m] at the end, but seems like you reached a character limit \o/ 20:03 MTDiscord katp: there is a "game time" value in the env_meta.txt file in a world 20:03 katp32 awesome, thanks 20:03 MTDiscord I think it's measured in seconds, and represents the total amount of time that the game world has been "ticking" 20:04 katp32 out of curiosity, is it possible for a mod to access that value? 20:04 MTDiscord though it's obviously not necessarily a good measure of how much time has "passed" in a particular locality since time is "frozen" when mapblocks are unloaded. 20:04 MTDiscord minetest.get_gametime() 20:04 katp32 great 20:04 MTDiscord It's an integer value so if you need sub-second precision then you'll have to do something else, like using the dtime parameter in globalsteps or something. 20:05 katp32 sorry, to clarify, is this per-world or per-mapblock, and how does it behave if the server is up but there are no players on it? 20:06 MTDiscord it's per-world, and it keeps ticking regardless of players 20:06 katp32 ah, great 20:06 MTDiscord It looks like most games using irrlicht had to fork it ... 20:06 MTDiscord In Deiban package, supertux not even depends on it... nor it shows anything related to it from a first look 20:07 katp32 well, making a completely generalized game engine is difficult, and introduces inefficiency 20:07 VanessaE so just bundle the damn thing with the engine. make it easy on us old pharts :) 20:09 MTDiscord That's the kind of use-case that submodules are designed for. 20:09 ar10ch my boy released a mod yesterday but we dont know how to use github or the forum so no idea if he's done it right. 20:11 VanessaE Warr1024: like I said, submodules suck. but in all fairness they suck way less than not including it at all (and hence having no preconfigured, predictable location to #include it from) 20:11 MTDiscord > That's the kind of use-case that submodules are designed for. Agreed, but seems like this practice has a lot of hate rather than love. Regardless, having a local vendored copy makes sense to me, specially if there are changes linked. Anyways, it would be nice if a local checkout can be detected and used. 20:11 MTDiscord VanessaE: yeah, that's about how I feel about submodules too. They suck, but just less than most alternatives. 20:12 MTDiscord Taking from supertuxkart example, they have a copy, and from time to time update/sync with it. Their build always include it. If it is not a big of a coding hassle, detecting a local checkout in the lib folder would make a lot of sense and updating the REAMDE to clarify that would also be helpful. 20:14 MTDiscord With or without submodules, it would make compiling straightforward. The only advantage I see over a submodule is that you can be more clear as of what dependency version a given checkout refers to. If a user reports an issue, it could be harder to identify it. Like, what irrlichtmt + minetest engine commits are involved. 20:15 ar10ch it's called mario_parkour on github if anyones interested 20:15 MTDiscord Harder without the submodules I mean. 20:15 MTDiscord ar10ch: do you have a Github link? 20:15 ar10ch hold on - its not my account but ill find it 20:16 MTDiscord I suggest you post it on the Games sub-forum as well! The topic could help you share it with the community - https://forum.minetest.net 20:16 ar10ch Ronaldo: https://github.com/kingowainturbo/minetest-mario_parkour 20:17 ar10ch Ronaldo: Ok thanks - Is that here or on mt forum? 20:17 ar10ch oh sorry i saw it 20:18 VanessaE the textures appear to be ripped directly from a commercial Mario property.. 20:18 ar10ch He put it in WIP because there was no buton for new topic in mod releases 20:18 VanessaE (some do anyway) 20:19 ar10ch *button 20:24 MTDiscord Katp32, either builtin or mtg has a command for it, i dont recall which atm 20:25 katp32 a command for what? 20:25 MTDiscord World days 20:25 katp32 oh, there's /days, yeah 20:26 MTDiscord ar10ch: thanks for sharing! 20:27 * VanessaE wanders off, kicks some dust at irrlicht :P 20:27 ar10ch :) 21:06 DS-minetest is it just me or are checkboxes in the main menu no longer working? 21:30 doubtman I have a mouse with side button 21:30 doubtman How can I bound them to some action on minetest? 21:31 doubtman Side buttons 21:31 doubtman Sometimes called mouse 4 and mouse 5 21:36 doubtman Is it possible to bind these buttons for actions like run, dig, toggle map? 21:37 ar10ch maybe in minetest.conf ? 21:38 ar10ch if u know the key binding for the mouse 21:40 doubtman How can I know it? 21:43 sfan5 http://irrlicht.sourceforge.net/docu/namespaceirr.html#a54da2a0e231901735e3da1b0edf72eb3 here's the list 21:44 sfan5 could be KEY_XBUTTON1 or KEY_XBUTTON2 but not sure 22:43 ar10ch Hmm .. I did a mod a while ago (unreleased) but I couldn't satisfy the rotation issue. My "magic wand" places a mts, how do I do get the direction in which the schem is placed to match, say the facedir? 22:46 MTDiscord https://github.com/minetest/minetest/blob/master/doc/lua_api.txt#L5540 see rotation 22:48 ar10ch static rotation ? 22:49 ar10ch I can write in the rotation but it's not what a player wold want 22:50 ar10ch *would 22:54 ar10ch it's some kind of dir to facedir thing (i think) - ill have to try and remember where ive seen something similar 23:10 ar10ch ok i could write in rotations and say something like: "cobble / acacia house - north facing door", or "stone_brick / pine house - west facing door". If there's no better way. 23:40 ar10ch nvm i must be a genius for thinking of something nobody has done before :) 23:41 DivideByZero Shaders don't seem to be working in the latest git dev-build. 23:41 DivideByZero They are enabled in options. 23:50 DivideByZero with shaders off the canvas of NPC and mobs are white, with it enabled the canvas is black, but the graphic textures do not display. 23:51 DivideByZero Well most of them dont. I see about one texture per NPC or mob, displaying, and the rest of the canvas is black. 23:52 DivideByZero Is this a known issue? Have I somehow done a bad build?