Time Nick Message 00:00 Menchers would you remove the infinite water behavior to make pumps more useful? 00:00 Menchers not much need for a pump now, can carry two buckets to the top of a mountain and have infinite water up there 00:02 Gustavo6046 oh dear sorry my network's being a bit nuts today, what was the last message that sent? 00:02 Gustavo6046 I may take this opportunity to neat up a little on the words that I wrote so ADHD-ly prior 00:02 Gustavo6046 oh, pumps would be more like parts of other machines 00:03 Gustavo6046 could be used, say, in a water tower 00:03 Gustavo6046 so this water tower would have a pump as an energy input, and an adjustable flowrate canal with water turbines as an energy output 00:03 Gustavo6046 it'd essentially be an energy buffer 00:03 Gustavo6046 and heavier liquids are harder to pump up but produce more energy as they go down 00:03 Gustavo6046 so this water tower would have (A) a pump component, (B) a water turbine component, (C) possibly components that convert between this mechanical energy and whatever other kinds of energy may be desired by the player 00:04 Gustavo6046 and of course the water reservoir at the top might be expandable, idk how that'd work 00:04 Gustavo6046 yeah maybe the hydraulics / fluid machine simulation part would have to be separate 00:04 Gustavo6046 but, say you have a nuclear reactor producing massive amounts of energy with each run, but it's not producing consistently 00:04 Gustavo6046 you can plug up this water tower buffer to it, with a big enough water reservoir 00:04 Gustavo6046 maybe up to a lake of finite fluid, idk 00:04 Gustavo6046 it'd just spin the pump super fast and fill the reservoir about halfway or 3/4ths of the way with each run of the reactor 00:04 Gustavo6046 but then you could adjust how fast you want water to flow back down to convert into a steady outflow of energy 00:04 Gustavo6046 this could be made Technic compatible! 00:04 Gustavo6046 and if it goes well, it could potentially replace Pipeworks or even Technic itself as you can have an incentive to research new machine components and actually do schematics and stuff 00:19 MTDiscord can a minetest server be re-built from source in-place? It is currently a 5.4 server, can I rebuild in place? 00:33 Menchers rebuild in place? 00:33 Menchers you mean without a restart? no 00:33 Menchers Gustavo6046, I do really like the idea of small machines that can be composed together into a larger more useful machine :D 00:35 MTDiscord I mean rebuild minetest from source without having to build elsewhere and then copy the files 00:53 Gustavo6046 :D 00:54 Gustavo6046 Menchers: if it were to replace technic it'd "provide technic" and the existing nodes would be replaced by "default component trees" 00:54 Gustavo6046 same goes to Pipeworks 00:54 Gustavo6046 although the latter could break the finite fluid thing I was thinking about too 00:54 Gustavo6046 and also being able to only push liquids so far before needing more pumps 00:54 Gustavo6046 (although that could be another screw pump advantage - they can extend arbitrarily provided enough energy to move their weight) 00:55 Menchers MisterE, you can definitely do that 00:56 Menchers you can set RUN_IN_PLACE in cmake's configuration 00:56 Menchers run the server directly from the source tree 00:56 MTDiscord I know... but if it has already been built... 00:56 Menchers then you can fetch updates, rebuild the server while it's running, then just restart it to deploy it in seconds 00:56 MTDiscord ok great! 00:56 Menchers did you get it from git? 00:56 MTDiscord yes 00:57 MTDiscord im on openbsd but I have good instructions for building from source taht I know work 01:29 Menchers openbsd! 01:29 Menchers yay 01:29 * Menchers uses the open beastie 01:32 Menchers huh it turns out my crash problem is a resurgance of the same problem I had before on the same server https://menche.us/files/minetest.txt 02:27 MTDiscord do 5.3 clients recieve dynamic media? 02:28 MTDiscord Menchers: if you are trying to build minetest on openbsd, here are the instructions: https://wiki.ircnow.org/index.php?n=Openbsd.Minetest 02:29 Menchers I have it built on openbsd :D 02:29 Menchers and it usually works 02:29 Menchers just having a weird crash issue lately 02:30 Menchers but in only happens on a specific server, very intermittently 02:31 MTDiscord no 02:32 MTDiscord oh too bad ok thx 02:33 MTDiscord that is indeed a weird error 02:34 MTDiscord that is indeed a weird error 04:20 MTDiscord if anyone is interested: my mod that adds dynamically sent and individual background music, based on environment: https://forum.minetest.net/viewtopic.php?f=9&t=27361&p=401190#p401190 04:20 MTDiscord only for servers, and only for 5.5 dev +, clients 5.4+ 05:56 MTDiscord i keep having 05:56 MTDiscord oops 08:43 Kiwi yoyo 08:43 Kiwi GUYS 08:43 Guest2368 ANYONE KNOWS MEE 08:43 Guest2368 f 08:43 Guest2368 f 08:43 Guest2368 f 08:44 Guest2368 whos shadow 08:44 Guest2368 I LOVE SHADW 08:56 Guest2368 sad 08:56 Guest2368 sorry 08:56 Guest2368 bye 13:17 MTDiscord hey, I just upgraded my server to 5.5 dev and some of my clients are timing out. Im sure its not the new mod I added, since I disabled it and it keeps happening. Any clue as to how to start debugging this? 13:27 macaw Hey. I do not have minetest on my machine right now, but the general advice would be to look at the logs, see where something reports Errors and consider whether they are the cause. 13:28 macaw If the clients are timing out, you may have to connect a client yourself (to the server) and look at the clients' logs. 13:41 celeron55 the fastest way to get an idea would be to download some different client versions your users use a lot and just try it out 13:41 celeron55 it's not likely the server logs show much in a timeout case 14:17 MTDiscord well they werent timing out when it was 5.4... hmm 15:33 MTDiscord ok I tried it with 5.4.1 which people said they were using... but I got no lag. Regardless, they continued to get lag. Notes about it: the lag disappeared when I teleported to them it reappeared when they logged off and then on 15:33 MTDiscord also Mineman said he cant join 15:33 MTDiscord I am switching back to stable-5.4 for now ? 15:36 MTDiscord if you have any concrete ideas about how to work this out then I would switch back 15:37 sfan5 you can enable verbose output and hope it contains something useful, but that doesn't help directly 15:39 sfan5 other than that it would be useful if you had a rough idea when the bug was introduced (does it happen on 5.5-dev from one month ago? what about two months ago?) 15:39 MTDiscord oh 15:39 MTDiscord thats a good idea 15:39 MTDiscord the problem is that I cant get lag 15:40 MTDiscord only me players do 15:40 MTDiscord *my 15:40 MTDiscord so they have to test it 15:45 MTDiscord sfan5: any suggestion for which commit to start at? 15:46 sfan5 you can specifically test before fd8a8501bc26dfca2a93d51000867b8592210040 15:50 MTDiscord ok building that commit's parent commit now 15:52 AristotIe in the linux terminal if I do `git clone --depth 1 https://github.com/minetest/minetest.git` I get minetest 5.5.0-dev, what would the url be to get minetest 5.4.0? 15:53 MTDiscord you should clone the whole git repository and then checkout the branch or commit you want to use 15:53 AristotIe so that would be... 15:54 AristotIe *I'm not very farmiliar with github* 15:55 MTDiscord 1) build mt irrlicht 2) $ git clone https://github.com/minetest/minetest.git 3) $ git checkout stable-5 or git checkout master<- for latest dev version 15:55 MTDiscord 4) build minetest 15:56 AristotIe ah, thanks 15:56 MTDiscord between 2 and 3, $ cd minetest 15:56 AristotIe yeah 16:01 AristotIe huh, when I do step 3, it says `error: pathspec 'stable-5' did not match any file(s) known to git` 16:04 MTDiscord please paste the commands you did in order 16:04 AristotIe well 16:05 AristotIe `git clone --depth 1 https://github.com/minetest/minetest.git` 16:05 AristotIe `cd minetest` 16:05 MTDiscord sfan5: Mineman says that he was not able to join, when the server was on latest 5.5dev. I just built 5.5dev from commit https://github.com/minetest/minetest/commit/ea250ff5c57301b6ea3e529c811484c743c1fde1 and mineman was able to join 16:06 MTDiscord Aristotle, im unsure about that --depth 1 16:06 sfan5 AristotIe: re-do the clone command with -b stable-5 (delete the folder first) 16:06 sfan5 when you have --depth 1 checking out other branches may not work 16:07 AristotIe MisterE: well, I was just trying that, so I'll continue with that and then if that doesn't work, I'll refer back here 16:08 AristotIe I assume it worked this time, it gave me these 2 lines: `Branch 'stable-5' set up to track remote branch 'stable-5' from 'origin'.` 16:09 AristotIe Switched to a new branch 'stable-5' 16:09 independent_ Where did this all go wrong? my server crashes when i set some routes! I need to compeltely re-interlock this stretch of track 16:11 independent_ I feel like XKCD #1084 16:11 ShadowBot https://github.com/minetest/minetest/issues/1084 -- Add InvRef::get/set_lists() by ShadowNinja 16:17 MTDiscord sfan5: I can confirm that that commit causes the bugs 16:17 MTDiscord no more lag problems 16:18 sfan5 I know how my commit is related to timeouts but lag problems are new to me 16:18 sfan5 can you be more specific how those manifest? 16:25 MTDiscord so i do not experience lag, but mec404 would join, then after a few minutes, lag would start to build up. Then he would time out 16:26 MTDiscord but strangely, when I teleported to him, his lag stopped completely 16:26 MTDiscord until he logged out and then on again when I was not on the server anymore 16:26 MTDiscord 2 other players were unable to join the game at all 16:27 MTDiscord but now that I reverted to the parent of that commit, everyone can join and has no more lag than normal 16:27 MTDiscord is there a way to disable the hand completely? 16:29 sfan5 the visual part or the "the player can interact with things" part? 16:31 MTDiscord you can overwrite the hand definition 16:31 MTDiscord or make a new tool and set that as the hand 16:31 MTDiscord well, to explain, im trying to make a player only be able use a items on_place, setting the hand inventory to the item works (and main to 0), but in the hand inventory, on_place is not called. if hand inventory is set to nothing, and main is set to 1, and the item added to main, the player can still rotate over to the hand even though it doesnt exist 16:34 sfan5 sounds like you're not setting the hotbar size 16:36 independent_ Mostly fixed now. 16:36 independent_ I hate it when my infastructure breaks down 16:37 MTDiscord thanks sfan5 16:39 MTDiscord any hope for fixing this bug soon? 16:39 MTDiscord https://cdn.discordapp.com/attachments/749727888659447960/896436763550183454/unknown.png 16:40 MTDiscord where attached players' models are blown up 16:41 MTDiscord attached is set to the same scale as the root 16:41 MTDiscord you can just set it to the proper one after attaching to work around it 16:42 MTDiscord yes and draconis does that but anotehr mod is doing something to conflich with it 16:43 MTDiscord it would be better if the bug was just fixed in the engine 16:43 sfan5 "fixing" it would break all mods that apply the workaround currently 16:44 MTDiscord heh ok 16:44 MTDiscord maybe in 6.0 16:44 MTDiscord tho even then maybe not 16:44 MTDiscord if your going to fix that, should fix models of ents being 10 units to small 16:45 MTDiscord yeah so scaling the model in blender would fix it 16:45 MTDiscord but as sfan5 said, break sompat 16:45 MTDiscord ? 16:45 MTDiscord *breaks compat 16:45 MTDiscord yeah so scaling the model in blender would fix it? 16:45 MTDiscord or using visual_sacle 16:45 MTDiscord *visual_scale 16:46 MTDiscord its easier to fix in blender if your using attachments so that the attachements dont need to be resized 16:46 rubenwardy for this, you typically have a transform per entity (position / rotation / scale) that applies to all children, and a per-entity adjustment that only applies to itself 16:47 MTDiscord so its not a bug 16:47 rubenwardy it's misleading and badly implemented 16:47 rubenwardy the feature should be implemented as a transformation matrix on the entity, rather than an object property. And then a separate way to scale without scaling children 16:48 rubenwardy it's possible to fix this without breaking compat by making a new property for just scaling the entity and not its children 16:48 rubenwardy network compat may be tricky. I guess you could calculate the old visual_size from the new transform and scale bet eh 16:48 rubenwardy *but 16:54 sfan5 @MisterE if you have time please recompile the latest dev with this patch http://sprunge.us/z6WDeh?diff and report back if the problems appear again 16:55 MTDiscord I dont know how to do that... could you give me a branch of minetest that I can check out? 16:57 sfan5 https://github.com/sfan5/minetest/tree/temporary 16:57 MTDiscord ok, ill bookmark it, probably will be able to get back to you in within 24 hrs 16:58 MTDiscord players are on the server rn and having fun and I dont want to stop that right now, but later tonight maybe 17:02 sfan5 sure, thanks 19:46 * Menchers grumbles discontentedly about all the stupid carriage returns in the irrlichtmt source code 19:50 * Menchers randomly wonders why the [REDACTED] irrlicht implements its own string type 19:51 sfan5 that makes it portable to platforms with no (complete) C++ stdlib 19:52 sfan5 I think the real reason might have been that in early times the C++ standard library (esp. containers and related stuff) was considered bad so people wrote their own 20:09 * Menchers adds a debug printf, wonders why it segfaults immediately now 20:28 Menchers oh, this object is null 20:28 Menchers why's it calling crap with it? why doesn't this crash on other platforms? 20:28 Menchers dafook 20:34 Menchers huh, the clouds are now textured with the font map texture 20:34 Menchers wat 20:38 sfan5 ?! 20:43 Menchers I am very stressed right now 20:43 Menchers this code is not good for my mental health 20:44 Menchers I will have more to ask about once I finish this debugging session… 20:46 Menchers *sigh* 20:46 Menchers the crash is always at irr::video::ITexture::getDriverType 20:46 Menchers line 256 20:46 Krock re-compile Irrlicht and Minetest. ensure that Minetest uses the correct library 20:47 Menchers E_DRIVER_TYPE getDriverType() const { return DriverType; }; 20:47 Krock i.e. correct ldd output paths 20:47 Menchers (irr::video::E_DRIVER_TYPE) DriverType = 20:47 Menchers dunno what's wrong with DriverType but trying to return it is apparently a segfault 20:50 Menchers running out of ideas 20:51 sfan5 recompile both like Krock said 20:51 Menchers I have 20:51 Menchers many times 20:53 Menchers what could cause returning DriverType from ITexture to segfault 20:53 Menchers it's not NULL 20:53 Menchers I checked 20:56 sfan5 did you check that Minetest is loading the exact library file you compiled? 20:57 Menchers yes 20:58 sfan5 how? 20:59 Menchers the changes I'm making to irrlicht are showing up when I run minetest in the debugger 20:59 Menchers and I can look at ldd 20:59 Menchers it confirms 20:59 sfan5 okay good enough 21:03 Krock "it's not NULL". means that it's accessing garbage values. uninitialized pointers. -Wextra or valgrind could help there 21:03 sfan5 there's one thing you might want to know: if you build Irrlicht as debug it's possibly broken 21:03 sfan5 by default 21:03 Menchers what 21:04 Menchers do I need to change something in the flags CMake sets for debug builds? 21:04 sfan5 yes 21:04 sfan5 if you didn't set anything you should have a release build 21:04 sfan5 (so not affected by this) 21:04 Menchers I set debug build 21:05 Menchers changed it to -O0 -g3 -ggdb 21:05 sfan5 try -DCMAKE_BUILD_TYPE=RelWithDebInfo 21:05 Menchers lldb's information is very unuseful otherwise 21:05 sfan5 the crash might just go away magically 21:05 Menchers ok 21:15 Menchers if it goes away magically I won't figure out what caused it though >:/ 21:16 Menchers anyways it still crashes, but the debugger shows assembly instead of C++ now 21:19 sfan5 huh okay so that wasn't it 22:17 Swift110-mobile hey akk 22:18 Swift110-mobile hey all i mean 22:18 Swift110-mobile whos playing right now? 22:27 MTDiscord I can... 22:27 MTDiscord what server? 22:42 erlehmann Menchers have you tried using asan or msan or ubsan? 22:43 Menchers last I checked, those didn't work on openbsd 22:43 Menchers can check again 22:43 Menchers right now I'm rebuilding irrlichtmt with g++ instead of clang++ 22:56 Menchers doesn't fix it 23:07 luizsab hello. Someone already use the alchemy mod? I can't find documentation 23:07 luizsab Or how potion mod do you can recommend? Is there some better? 23:15 Menchers trying the build with -fsanitize=undefined 23:17 erlehmann Menchers good luck 23:18 Menchers :D 23:20 * Menchers should probably also do memory sanitizer… but later 23:30 Menchers hm, right, openbsd doesn't seem to ship with the sanitizer runtime built 23:30 Menchers unsure why 23:31 Menchers :( 23:35 MTDiscord there are at least two alchemy mods