Time Nick Message 00:09 Dacmot After all that, the bug is still there with 5.6.1 so I don't think it related to MT... thanks for your help wsor! 00:11 MTDiscord np, low key curious what mod 00:11 Dacmot 3d_armor 00:12 MTDiscord fun, good luck 00:13 Dacmot Getting "Armor not initialized!" when loading a new game in whynot. Joining again and it's gone. 00:15 erle Dacmot for some reason both submodules and subtrees are strongly opposed by coredevs. i have never seen a good explanation why and do actually assume that the people opposing both have just not had problems that submodules or subtrees would solve. 00:16 erle meanwhile, the people who do have these problems are not coredevs 00:16 Dacmot Hehe. Thanks for the info erle 00:16 erle this is a simple explanation that requires no malice or incompetence (just not caring too much about casual contributors) 00:16 erle Dacmot well i urge you to ask yourself in #minetest-dev 00:17 erle after all i might be mistaken 00:17 erle but the question gets asked from time to time 00:17 Dacmot Meh. I'm not looking to do MT development. Just wanted to build an old version to test something. A minor inconvenience is all. 00:17 erle yes but it is the channel for engine dev 00:18 erle Dacmot IIRC the current proposal for example to merge irrlichtmt into minetest proper is to copy it over or so? instead of doing a subtree merge (a merge strategy invented to nest repositories) 00:18 erle as always, don't believe random ppl on the internet 00:18 erle you can also download the chatlogs from #minetest-dev and grep for “submodule” 00:18 Dacmot Haha. Ok, thanks! 00:19 erle and when i say coredevs i mean coredevs as a group. if they can not *agree* on submodules it does not mean that every one is a submodule hater 00:19 Dacmot I understand 00:21 erle i am greping right now 00:21 erle to figure out if the answer WAS actually provided 00:23 erle i found a chat from nov 6 00:24 erle Dacmot start reading here about submodules from people i think https://irc.minetest.net/minetest-dev/2021-11-06#i_5894230 00:25 erle okay this is funny 00:25 erle so from my logs i am the person who asks the most for irrlichtmt to be a submodule 00:25 erle sfan5 is the person who most often is against submodules (project does not build if you don't use submodule and other reasons) 00:25 erle but almost everyone else is either in favor of it or is like “i don't care, but people seem to hate submodules” 00:25 erle hahaha 00:26 erle i think that everyone hates submodules either was agreed upon outside the time that my chatlogs span or it is an illusion 00:26 erle like, surely some people dislike them 00:26 erle but they can not be as disliked as everyone who likes them is claiming, because not too many people openly argue against them? 00:28 erle so regarding the “not seen a good explanation”: i have seen explanations but i think they are bad (i.e. the arguments seem not too convincing enough even if i agreed with them) 00:28 erle but ultimately it does not matter 00:28 erle as i said, if the group can't agree, it does not happen 00:29 erle and i think for this reason i predict that subtree merge will not be the merge strategy used for irrlichtmt. instead, i guess it will be squash merged and git blame will then forever be wrong and git bisect will lead to a huge commit ^_^ 00:30 erle hmmm, now i wonder if there is a modified subtree merge that could rewrite minetest history to always provide the correct irrlichtmt hmmmmmmmmm 03:29 MTDiscord Does anyone know if there's a way (existing, already implemented, or at least already WIP) to get texturepack overrides.txt honored server-side? It seems like MT allows you to unpack a texture pack into a certain place to serve as a "server-side texture pack" and override all the texture images, but it doesn't honor the overrides.txt (not sent to client and/or not parsed). 08:34 rubenwardy Warr1024: you can either convert the texture pack into a mod or patch the engine 08:35 erle well the first thing would be to open an issue right 08:35 erle i was not aware of this 17:04 MTDiscord I actually tried doing it for rpg16 I think it worked? https://github.com/Lemente/rpg16_mod/blob/main/init.lua 17:08 MTDiscord since it is parsing the override.txt file I asume it could work with any texturepack 17:32 erle lemente you planning to generalize it? 17:35 MTDiscord I forgot I made it, so ... 17:50 MTDiscord I guess I could try dropping it into other texturepacks, see if it works, and provide it on contentDB with a How To. Was it what you meant by "generalize it"? 18:02 erle lemente you could make a mod that does this thing for all server-side texture packs? 18:02 erle lemente yes i meant that 18:11 MTDiscord I guess that would be possible, you wouldn't even need all the textures in the mod since the texturepack would already provide them (contrary to my mod which purpose was to completely replace the TP). I imagine the mod would need to be trusted though, and to know the TP folder. Is that possible? 18:22 erle not sure 18:22 erle elment then again you could load after all other mods 18:22 erle and just STEAL the node textures 19:20 erle new funny idea: i may be able to make a TGA image that when a user calls file(1) on it outputs its own representation to the terminal in ASCII art 19:21 muurkha that would be fantastic. did you find an arbitrary execution hole in file(1)? 19:22 erle no, i think it just prints arbitrary strings from the TGA file 19:22 erle if it prints newlines, i could do it 19:23 erle muurkha i ran file(1) on files in this repository https://github.com/lunapaint/tga-test-suite/ 19:27 erle ok newline does not work 19:30 erle ; file tombexcavator/rgb15.tga 19:30 erle tombexcavator/rgb15.tga: Targa image data - RGB 199 x 199 x 15 - HELLO\011WORLD! 0.0 19:34 muurkha you could format to 80 columns 19:35 erle yes but 19:35 erle first i read where the magic happens 19:36 erle i have to understand /usr/lib/file/magic.mgc it seems 19:36 erle why oh why is this a binary fail? :( 19:36 erle file lol 19:38 muurkha they started compiling it to make file faster 19:39 muurkha it's built from textual sources 19:40 muurkha file-5.38/magic/Magdir/images 19:41 muurkha # originally from jef@helios.ee.lbl.gov (Jef Poskanzer), 19:41 erle > # at 1, byte CoMapType must be 1 if ImgType is 1 or 9, 0 otherwise 19:41 erle hmmmmm 19:41 erle i wonder i wonder 19:42 erle this is horrible 19:42 erle but also a funny source for parser differentials 19:43 erle > # Compressed color-mapped data, using Huffman, Delta, and RLE. 4-pass quadtree- type process 19:43 erle what 19:43 erle i'm not doing this (and minetest is not doing it either) 19:44 erle one comment is literally “gimp can create big pictures!” 19:44 erle > # image height. 0 interpreted as 65536 19:44 erle thanks libmagic 19:46 erle yeah so libmagic reads the author name and a comment 19:46 erle > PNG [Portable Network Graphics, or "PNG's Not GIF"] images 19:46 erle hahahaha 19:47 erle let's see 19:48 erle oh it has only an IHDR parser, lame 21:18 MTDiscord @erle "you could load after all other mods" there are two issues with this: - I believe I would get the texture from the mod, not the texture pack (the texturepack loads after all mods) - the override.txt can add completely new textures to a node (like a different top and bottom) that's why I needed to make a mod from the override.txt in the first place, so another mod (texture generator by LMD) would also modify the textures in 21:18 MTDiscord override.txt now that I think of it, maybe that mod should have been handling it in the first place 21:19 erle oh i see 21:22 MTDiscord I thought it was last loaded mod whose textures won. 21:22 MTDiscord So you could put the textures either in the mod, or in the server texture pack space, either should work 21:27 MTDiscord yes, but the TP will supersede every mod afaik 21:49 MinetestBot 02[git] 04jiaowobaba02 -> 03minetest/minetest: Translated using Weblate (Chinese (Simplified)) 13d62abbc https://github.com/minetest/minetest/commit/d62abbc9381a59151bda6e33407855610abc71ea (152023-10-20T20:54:04Z) 21:49 MinetestBot 02[git] 04Gaemy34 -> 03minetest/minetest: Translated using Weblate (Esperanto) 1317e0ec2 https://github.com/minetest/minetest/commit/17e0ec27ebd43f0177f79e3e77e918a653092f64 (152023-10-20T20:54:04Z) 21:49 MinetestBot 02[git] 04Uwuewsky -> 03minetest/minetest: Translated using Weblate (Russian) 1377f2c94 https://github.com/minetest/minetest/commit/77f2c94395f19781570a94a45d0ec2112cb00d74 (152023-10-20T20:54:04Z) 21:50 MinetestBot 02[git] 04Linerly -> 03minetest/minetest: Translated using Weblate (Indonesian) 1328e06f7 https://github.com/minetest/minetest/commit/28e06f7d9cc0e79ab06dc3109aa448049196bf3e (152023-10-20T20:54:04Z) 21:50 MinetestBot 02[git] (16 newer commits not shown) 22:27 erle I HAVE BEEN VISITED BY THE GHOST OF THE EXTREMELY 1990s UX 22:27 erle (the new settings dialog) 22:28 MTDiscord LOL 22:29 erle it's the mystery meat navigation 22:29 erle technically, it's excellent 22:30 erle i just think that some UX person should probably look at it 22:31 erle (i am not one and could not do it better, i am just pointing out issues) 22:31 erle ROllerozxa the funniest thing was that i told a friend i had difficulties finding the settings dialog 22:31 erle and she was like “yeah, i KNOW it changed and i still have difficulties finding it sometimes” 22:31 erle (because it is unlabeled now) 22:32 MTDiscord Settings UX hasn't changed for me, it's still just vi ~/.minetest/minetest.conf, same as it aways has been... 22:32 MTDiscord the new settings button is a bit small, if anything... 22:32 erle ROllerozxa it is small and unlabeled, so it is both a smaller click target and harder to find 22:32 erle and also i have no idea what the icon is supposed to represent 22:33 erle it can't be gears, they are too smooth 22:33 MTDiscord Ain't gonna say that I dislike the new settings dialog per se, I suppose it's more just that the settings I find myself needing to change stopped being the ones those things ever helped with a while ago... 22:33 MTDiscord I thought the whole "settings button" thing was just supposed to be a stopgap, and that the rest of the main menu needed to be changed at some point anyway. 22:34 erle yeah but why make it *worse* than before? i mean you can easily have a label AND the icon there in the corner 22:34 erle the space is there 22:34 MTDiscord nothing's as permanent as a temporary stopgap measure 22:34 MTDiscord erle: do you think a regular button with a text label would work better for the settings dialog? 22:34 MTDiscord Well, the eggs have been broken, I guess ... now I just hope the omelettes arrive someday... 22:35 MTDiscord hammer and/or screwdriver might make another "settings" icon alternative, which might hold up better under scaling ... worth experimenting with, I guess. 22:35 erle ROllerozxa i personally think the styles ”text besides icon” or “icon above text” are the best ones, because they provide both an icon and a label. and they give the biggest click target. 22:35 erle Warr1024 it's not about the icon 22:36 MTDiscord Heh, yeah, that button looks less stopgappy than it did before, but it's a bit TOO unobtrusive. I feel like I need to be obtrused a bit more about where I can change my settings. 22:37 erle GNOME had this hilarious UX study where they found out that when they switched away from menu bars to “burger menus” (the three horizontal bars that scream “the person who made this did not ask a UX person for input”) … people did not find stuff anymore. 22:37 erle so they were like “maybe if we change the icon of the button” 22:37 MTDiscord ahh, GNOME's hamburger menu of doom 22:37 erle yeah 22:37 erle the reason why it is bad UX is because it invites people to no longer hierarchically order menu items 22:38 erle just stuff everything into it 22:38 erle though i think at some point apps started having 2 of them 22:38 MTDiscord The correct way to make a hamburger menu is by using the actual hamburger emoji instead of just those lines, so that all the people who would be complaining about your bad UX are instead complaining that the cheese is under the meat instead of on top. 22:38 erle this reminds me of the flash game that mcdonalds had years ago 22:38 MTDiscord mmm hamburger 22:38 erle where people managed to insert references to image files 22:39 erle so they tried to make the ”McMyLittlePony” win 22:39 erle and also they managed to have a burger with chicken stuff as the buns 22:39 erle and the McArte (which was black and white and had the logo of the french-german TV station ARTE in the corner) 22:39 MTDiscord why do they call it hamburger menus if they don't offer any hamburgers 22:39 erle i don't know, why do they call it photoshop if i can't buy pictures? 22:40 MTDiscord I've been to a few restaurants in my day and I've never seen one where you have to get the hamburger first to be able to access the menu. 22:40 erle anyway, i only have made websites 22:41 erle can someone of you test if they can bind a key the character that ISO_Level3_Shift + any other key produce in minetest? 22:41 erle it's the modifier key that is above the left shift key 22:42 MTDiscord That's Caps Lock on my keyboard. 22:42 MTDiscord keyboard(s) I guess I should say 22:42 MTDiscord Does it matter which layout? 22:43 erle Neo2 22:43 erle but in general, i think some refucktoring might have broken some modifier keys 22:44 erle https://github.com/minetest/minetest/issues/13899 see here for symptoms 22:44 MTDiscord Thankfully it seems that the apostrophe key that was broken in chat on OSX is now not broken. If they have to unrefucktor I hope they don't regress that. 22:44 erle lol 22:44 erle Warr1024 what is ISO_Level3_Shift on your keyboard anyway? 22:45 erle i'll try any modifier keys and see if anyone of them is registered separately 22:45 MTDiscord I'm on QWERTY USA keyboards, the only modifiers I know of are Shift, Ctrl, Alt, and Super, and Command and Option on Mac which are weird and I don't know how they work anyway... 22:46 MTDiscord ISO_Level3_Shift really sounds like something I don't have the necessary security clearance for. 22:46 erle well i can bind a key to ISO_Level3_Shift which should not be possible 22:46 erle it would be like binding a key to command or option 22:47 MTDiscord I've got Fn keys that are supposed to be normally unbindable, I think. 22:47 erle yeah but i mean modifier keys in particular 22:47 erle because i need to press a modifier key to input the + 22:48 MTDiscord I've got 2 keyboards where + is shift =. In theory there's a numpad+ on full-width keyboards but I gotn't one of those in fronta me right now. 22:48 erle the opengl 1.4 rendering seems still not broken. i think maybe it is because ppl are more occupied with breaking OGLES 2 for smartphones :P 22:48 erle Warr1024 so can you input that + then in minetest or not? 22:48 MTDiscord In my experience numpad + and shift= are treated as different things for "keybind" purposes; even though they input the same symbol, for purposes like games, usually I'm allowed to bind different functions. 22:49 erle well for minetest hat is now the case 22:49 MTDiscord I have no issues entering a + in chat. I have not tried using + as a keybinding separate from =, but since I'm able to bind shift for game control purposes, I expect that NOT to work. 22:50 erle yeah, i can input in in chat too 22:50 erle ROllerozxa i think the whole unbundling minetest_game thing has one problem. old versions of minetest game crash with new versions of minetest. 22:51 erle because minetest changes API to be incompatible with older versions of itself 23:21 rubenwardy Unfortunately, formspecs don't provide a nice way to have an icon and a label in one button 23:21 rubenwardy Closest is making a composite button 23:23 erle rubenwardy so what is the icon supposed to be anyway? 23:23 rubenwardy Cogs 23:23 erle oh lol 23:23 erle i see 23:24 rubenwardy A single cog is the more standard icon 23:24 erle well then for plain text it could always be U+2699 ⚙ 23:37 erle rubenwardy so where is the old settings dialog now? gone? 23:38 rubenwardy It's gone 23:41 MTDiscord (reduced to atoms)