Time Nick Message 05:22 kaeza sofar, re: game#1579, can you skip the realtime updates to homes_file and do it only at shutdown? 05:22 ShadowBot https://github.com/minetest/minetest_game/issues/1579 -- Migrate sethome mod to player attributes. by sofar 05:23 kaeza in sethome.set I mean 05:25 kaeza I'd personally prefer if mods started using the modname:whatever notation for everything that requires keys 05:25 kaeza e.g. sethome:home for the attribute in this case 05:26 kaeza (and FWIW, modname.whatever for settings) 05:28 sofar Since the list will only ever shrink, and shutdown may omit calling on_shutdown .. this is just simpler and almost more reliable 05:28 sofar I could care less about the prefix, I'm just the first one using it 05:29 sofar so, if people agree which one of [_:.] we use then I'll apply it 05:29 kaeza fine with me FWIW 05:31 sofar game#1608 05:31 ShadowBot https://github.com/minetest/minetest_game/issues/1608 -- Choose player attribute modname prefix character. 05:31 sofar nore: paramat: sfan5: rubenwardy ^^ 05:38 kaeza other related things are form names, meta keys, etc. 05:39 kaeza there seems to be default:chest_locked and default:book, but beds_form 05:39 kaeza that's a minor point though 05:40 kaeza I guess it's too late to change "secret" to "default:key_secret" or something 05:42 sofar they're often derived from lua names 05:42 sofar and those can't have : in them 06:46 celeron55 "I hope I am contacting the correct person here. I have been doing some testing with Minetest in a school I work in and I am very impressed so far. The only question I have is, is there a way I can secure this for children? Basically I only want them to access the single player and maybe an internal server. I know this isn’t what it was originally designed for however the school are really keen to deploy it network wide but, ... 06:46 celeron55 ... having publically accessible servers is a Child protection risk. If this is not possible do you have a list of servers or ports I am able to deny access to? To stop publically accessible servers functioning?" 06:46 celeron55 what's our reply to questions like this? 06:49 VanessaE imho, tell them this is outside of minetest's purpose, block external access at their firewall, save for a whitelist 06:50 VanessaE it's up to him to decide what isn't acceptable (or what is) 06:51 VanessaE (I'm a strong supporter of freedom of association, so if he wants to restrict that, let him do so on his own) 06:53 celeron55 well i think there's nothing wrong in what they're wanting to do 06:54 celeron55 someone's going to blame the school if the kids go on public servers and see things someone doesn't want them to see 06:55 VanessaE that's fair, but there have been "didn't go far enough" situations as well 06:56 VanessaE some dickhead somewhere will blame MT if it doesn't block things "just right" 06:57 celeron55 yes, we should certainly try to avoid that 06:58 celeron55 that's actually a very good reason for why we might want to not really support this 06:58 VanessaE of course, we could just blame nerzhul ;) 06:59 nerzhul ? 07:00 VanessaE nerzhul: just nod, smile, and say "okay" :) (or check the logs for 15 mins ago) 07:00 nerzhul i'm reading it , but i just wake up it's rude :p 07:00 VanessaE heh just kidding, nerzhul :) 07:02 sofar celeron55: I would actually suggest removing the multiplayer tab entirely instead 07:02 sofar or replacing it with a custom single hardcoded server form only 07:03 VanessaE the teacher (or whatever his/her title) could perhaps set up a proxy that captures the cURL request from MT and filters it? 07:03 VanessaE (I hesitate to lump that in with firewall rules) 07:04 sofar I would actually avoid external-to-mt-client solutions and patch the client instead so it does what they want 07:04 nerzhul just a mt.conf switch could be sufficient ? 07:04 sofar nerzhul: no, because the kids can modify mt.conf 07:04 sofar best to provide a client that has the abilities entirely removed 07:04 nerzhul just change rights to root root 644 07:05 sofar ~/.minetest/minetest.conf 07:05 nerzhul this is not a problem 07:05 VanessaE nerzhul: what stops the kid from launching MT with a different config? 07:05 sofar minetest --config some/other/conf 07:05 nerzhul path != rights 07:05 nerzhul ok 07:05 nerzhul but what is the problem in fact ? 07:06 sofar like I said, the best way is to hard code the client to only allow connections to a certain server 07:06 sofar so, patch the client 07:06 sofar anything else is a bandaid 07:06 VanessaE nerzhul: the teacher is afraid his students will end up finding pr0n. 07:06 nerzhul in mt ? lol 07:06 sofar VanessaE: s/will end up finding pr0n./ parents' will sue the school/ 07:07 VanessaE sofar: I was being sarcastic. 07:07 sofar I know, but, it helps to rule out kids being the problem 07:07 sofar the real problem is the lawyers 07:09 VanessaE I'd say the real problem is litigious parents, not so much the lawyers they hire 07:09 celeron55 then there's an email from someone from french asking if i or some other minetest developer could come to some "graphics labbor 2017" event in rennes, france to "share your experience about the minetest development, and at the same time, show us your others creations." and they are checking if they can pay for the trip 07:09 celeron55 france* 07:10 celeron55 i'm not going, but if some core developer or contributor wants to, i can forward this 07:10 nerzhul sorry i'm french but not interested, i'm not very aware on graphic parts 07:11 VanessaE celeron55: maybe you can talk Calinou into it ;) 07:11 nerzhul or nroe 07:11 nerzhul nore 07:11 VanessaE oh right 07:11 VanessaE or kilbith :) 07:13 celeron55 i am convinced this email comes from france because i can't really understand what half of it exactly means 07:13 VanessaE celeron55: that's a pretty nice offer though, in all seriousness. 07:15 celeron55 it is 07:40 paramat heh talking about MT at a graphics event =] sharing the experience of chaos development 08:08 celeron55 well they say it's a game developer event 08:08 celeron55 dunno 08:16 juhdanad If someone would like to go, I would happily explain how node colorization works. 08:25 paramat http://rgba.fr/rgba-invited-at-grafik-labor-in-rennes-france/ might be this 08:25 paramat "Grafik Labor is a new conference dedicated to graphic and visual creation using Free and open-source softwares and tools." 08:25 Calinou VanessaE: sorry, I can't, but thanks for the offer :) 08:25 Calinou celeron55: yeah, they asked on #minetest-fr 08:42 celeron55 fyi, here's my response to the earlier email with the school whitelisting question: 08:42 celeron55 Minetest isn't really designed to do that, but depending on what exactly you need, you could for example configure the public server list to be disabled or for example edit the main menu to not allow entering any other servers than a single one. Or, really, anything. The menu is a very flexible system for anyone with a bit of programming knowledge. 08:42 celeron55 The main menu lives in a bunch of Lua files, no need to even 08:42 celeron55 recompile: https://github.com/minetest/minetest/tree/master/builtin/mainmenu 08:43 celeron55 i.e. making a custom version for schools is easy, we don't need to stress about it 09:00 paramat please could someone review #4967 ? 09:00 ShadowBot https://github.com/minetest/minetest/issues/4967 -- New bulk node light update by juhdanad 09:26 paramat celeron55 perhaps you should state a legal disclaimer in your response? 09:29 celeron55 i asked about why the other ones are inviting minetest to a graphics-related event; this is the answer: "It seems I make a wrong express, this is not an event ONLY on graphic software, but also on the free web and video game development (and all the job about the video games). So, the event is about the developers, designer, sound designer, tester etc." 09:30 celeron55 so it's actually some kind of... well i don't even know, everything apparently 09:31 celeron55 we need some kind of a stand-up comedian who understands minetest's development history and can go to all events 09:32 celeron55 then it doesn't even matter what the event is about 09:34 celeron55 paramat: meh, there's enough legal disclaimer is in minetest's license already 09:34 celeron55 -is 09:34 paramat ok good 09:35 paramat hehe yes MT dev is a comedy 10:13 est31 servers.minetest.net -> 0.0.0.0 in hosts.conf 10:13 est31 very easy trick. works without a firewall 10:14 est31 ofc manual entry still possible 15:23 kilbith paramat should stop being a fucking hypocrite seriously 15:24 kilbith in one of his comment he invited Krock to downvote a PR and ignore the other developer opinions because of his "good sense", but then he silently deleted that part few hours later 15:24 kilbith sorry to be rude but that shit he regularly does must stop 15:24 kilbith be a real man and assume the consequences of what you're saying 15:24 kilbith ^ Zeno` could second it I think 15:30 Zeno` he said that? 15:30 kilbith yeah 15:30 Zeno` where? 15:30 kilbith I did not take a screenshot this time 15:31 kilbith in that comment : https://github.com/minetest/minetest/pull/4929#issuecomment-284974704 15:33 kilbith sometimes I dream to take this chickenshit by the collar and eject him through the door 15:33 kilbith he's not only incompetent, but hypocrite 15:34 Zeno` I really don't want to turn email notifications back on but maybe it's necessary :( 15:34 kilbith and impopular at the utmost degree 15:40 kilbith I'm willing to bet he'll say I'm exaggerating you'll see 15:44 Zeno` a comment like that would not be acceptable 15:44 Zeno` I can't see it, but if paramat wrote in a PR telling other people to downvote then that is BS 15:45 kilbith thing is, he frequently posts offenses and then edit like a coward 15:51 est31 This strategy done in a few exceptional cases is okay, but if you adopt it as habit its not okay. Your habit should instead be to read through the comment before you press the button. 15:53 nrzkt i also noticed paramat edited his comments to remove some parts after someone answer to it :( 15:53 kilbith especially when you're core-dev, otherwise what idea of MT you're leaving to the exterior... 15:54 sfan5 kilbith: the email notification does not include anything like that, must've been edited later 15:54 est31 well yeah image is important, but if you start kicking out people out of these reasons, you'll soon be empty. Nobody is perfect. 15:55 kilbith sfan5: yes, he may have added that part later, then removed it 15:55 kilbith damn on me to do not have a screenshot on hand 15:58 kilbith est31: it's not only that btw, it's not fun to contribute with him in the team; everytime it's about graphics he makes a fuss to make it difficult to merge 16:34 kilbith https://github.com/minetest/minetest/pull/4929#issuecomment-285091091 16:34 kilbith why the hell he's not coming on IRC 16:35 VanessaE he has said in the past that he finds IRC intimidating or difficult to talk on when there's a conflict. 16:35 kilbith like you said, chickenshit 16:35 * VanessaE shrugs 16:36 kilbith people with no balls these days... 16:37 kilbith meh, always the same excuses, gonna ignoring it 17:17 paramat kilbith i'm here, but i haven't joined IRC because you expect me too. often when someone demands i join IRC for some reason i don't 17:18 kilbith glad to see you finally came here 17:19 kilbith I don't intend to bite you, just demand to be honest and true with people on github 17:19 * sofar politely waves 17:20 paramat note that you responded to me on IRC instead of in the thread, likewise i will respond to you wherever i want to 17:20 kilbith it was offtopic, I didn't want to pollute that PR 17:23 paramat i'm happy to be honest and true about the unreasonable things i sometimes write, i don't delete them in the hope no-one sees them, as i have explained in the thread 17:24 kilbith I'll likely end up sending an e-mail to the github team to forbid editing comments after an hour 17:24 paramat the editing is not done through cowardice 17:24 paramat that would be silly 17:25 paramat editing is essential and useful for many reasons 17:26 kilbith yeah, to avoid conflicts for example ? 17:27 kilbith I've seen your numerous attacks toward e.g. hmmmm or Zeno` 17:28 paramat i don't attack hmmmm much. arguments happen in dev, that's what it's like 17:29 sfan5 kilbith: the github team wouldn't fulfil that request 17:29 kilbith what I hate is people being sugar with you on IRC and then they stab you on github 17:29 kilbith what an hypocrisy 17:31 paramat i'm happy to be unpleasant to you on IRC if you want :] 17:31 kilbith I prefer you go unpleasant with me on IRC but then it stops there, not on github 17:31 paramat i'm often unpleasant to people on IRC, uunfortunately 17:32 kilbith if you want a mudfight, do it here, but only here 17:32 Zeno` paramat, did you tell Krock to downvote a PR? 17:33 paramat 'tell' is not the word 17:35 paramat "Krock, i think you should stick with your opposition, according to your good sense, and not worry about what other's think, they may be wrong" that's the best i can remember 17:35 kilbith I confirm that was something like that, maybe a little more directive 17:35 paramat as he's clearly opposed to the concept as you can see from the comments 17:36 paramat although now i can see the PR is actually a lot more reasonable than i first thought 17:37 kilbith meanwhile I must leave now, bye 17:37 est31 I'll leave too 17:37 est31 bye 17:37 paramat meant as encouragement, but then i saw that trying to influence someone's vote is probably unreasonable 17:38 * est31 is gone. 17:38 * VanessaE remains, just for spite :P 17:42 paramat Krock sorry for that comment 17:42 Krock please specify 17:46 paramat hehe 17:46 Krock ah, *sees logs* 17:46 paramat see logs for this channel 17:47 Krock looks like a potential flamewar 17:48 Krock Zeno`, I'm the manipulating person here. Upvotes aren't given away on demand ;) 17:49 Krock *downvotes aswell 17:52 Krock is it my lossy brain or did I actually never hear this sentence about sticking with my opposition? 18:31 paramat Krock i edited my comment in the thread to remove that, so you might have not seen it 18:32 Krock ah, okay. 18:39 Fixer can someone write a simple mod that makes lua OOM? 18:40 VanessaE Fixer: why not just declare a table and fill it with really large elements? 18:41 sofar Fixer: a = {}; for i = 1,1000000 do a[i] = table.copy(minetest.registered_nodes); end 18:41 kilbith an infinite loop triggers a OOM quite fast 18:41 VanessaE foo = {} for i in 1,1000000 do table.insert(foo, "meh") end bar = {} for i i 18:41 VanessaE damn it sofar 18:41 VanessaE ninja'd :P 18:42 VanessaE (and yeah, I know. = instead of in. whatever :P ) 18:42 sofar please note that most OOM errors you see in lua are actually stack overflows 18:42 sofar but, that's a side-effect 18:43 VanessaE sofar: that raises the question, why wouldn't such a stack overflow be propagated to the error handling code before it could result in an OOM? 18:43 Fixer i need one that goes past lua mem limit, but don't use up all my 8gb 18:45 sofar Fixer: lua memory limit is less than your ram anyway 18:46 sofar VanessaE: a stack overflow is technically a segmentatio fault, so it follows a different error handling path 18:46 VanessaE hm 18:46 sofar that's why in C it also is handled differently 18:46 VanessaE gotchya 18:46 Fixer sofar: yes, but they introduced some new LJ_GC64 mode, I'm trying it out https://github.com/LuaJIT/LuaJIT/issues/25#issuecomment-183660706 18:46 sofar a stack overflow in C is not a -ENOMEM for instance 18:47 VanessaE (I never really connected stack overflow to a segfault before, but it makes sense) 18:48 sofar you're violating the segmentation of the stack segment 18:48 sofar by attempting to grow the stack beyond the limits of the stack segment :) 18:48 VanessaE right, that's what I figured 18:49 VanessaE I'm just used to old systems where the stack was not just a segment of memory, but restricted - an overflow in those instances overwrites the other end of the stack instead of exceeding its address space 18:49 VanessaE (or trying to exceed) 18:59 paramat Fixer this old version of riverdev might OOM, but it's not a simple mod https://github.com/paramat/riverdev/tree/9b34f87c9a35d6d7f2251e5ec8064e1b81d7164f 19:00 sofar VanessaE: gcc protects the stack from overflowing 19:00 paramat 10 3D noises =/ 19:05 getbiomeinfo *,.....,* 19:06 getbiomeinfo sofar: thank you 19:06 getbiomeinfo great results 19:06 getbiomeinfo I've used "a = {}; for i = 1,1500000 do a[i] = table.copy(minetest.registered_nodes); end" 19:06 getbiomeinfo gave instant OOM on loading with usual minetest 19:06 getbiomeinfo but now 19:06 getbiomeinfo i load this one with LJ_GC64 mode and it loads! 19:08 getbiomeinfo i will add lua mem info to that, to confirm 19:09 getbiomeinfo i will look more into that 19:19 sfan5 est31: you don't seem to understand 19:19 sfan5 "Fix various problems with sneaking" is not a series of changes 19:19 sfan5 it is a single change 19:20 sfan5 fall damage avoiding cannot be prevented without the 0.4 -> 0.29 change 19:20 sfan5 sneak-jumps cannot be fixed without doing it 19:21 sfan5 all of the changes non-bugfixing changes we're pretty much required to fix the bugs 19:21 sfan5 I did not choose to break sneak ladders 19:22 sfan5 it just came with fixing the other awful problems 19:23 paramat btw i tested and +1 #4967 please can anyone review? very important PR 19:23 ShadowBot https://github.com/minetest/minetest/issues/4967 -- New bulk node light update by juhdanad 19:47 getbiomeinfo yep 19:47 getbiomeinfo lua oom mem usage was at nearly 5GB... without hitting any limits 19:51 getbiomeinfo oom word is not needed in sentence above 19:53 getbiomeinfo too bad it is crashing :( 20:26 getbiomeinfo ShadowNinja: https://github.com/minetest/minetest/issues/2988#issuecomment-285157827 20:48 juhdanad Hi all! 20:48 juhdanad I need the help of somebody who worked on clent side modding. 20:49 juhdanad What is the difference between ScriptApi and ModApi? 20:57 sofar red-001: nerzhul: ^^ 20:58 red-001 iirc one is the interface from c++ to lua for callbacks and the other is from lua to c++ 21:02 red-001 yeah script api is c++ to lua and modapi is lua to c++ 21:02 red-001 what are you planing to work on juhdanad ? 21:03 juhdanad I tried to implement get_pointed_thing in lua but then I realized that there are no object references on the client. 21:04 juhdanad And I did not understand why were there so many different ModApi and ScriptApi classes. 21:08 VanessaE sofar: if you have a moment? GreenDimond there needs a quick review of a mt_game PR he's submitted. 21:08 VanessaE https://github.com/minetest/minetest_game/pull/1609 21:10 VanessaE (or paramat or whoever else wants to look) 21:12 VanessaE (he's going over to a proper IRC client) 21:12 sofar done 21:13 VanessaE thanks. 21:13 sofar I would just solve it outside of mtg, and in the mod 21:26 nerzhul juhdanad, ClientModApi is lighter than server because we don't need many things, but you are free to propose a PR on CSM branch. But as i said in a message on the PR i think we should merge it this weekend if no core dev has good reason as we have worked at to make it working with many securities and the API light but secure and we will add more callbacks depending on user needs 21:29 juhdanad nerzhul: I'm also waiting for the merge! Then #4421 may have chance to be merged. 21:29 ShadowBot https://github.com/minetest/minetest/issues/4421 -- Expose getPointedThing() as Raycast by juhdanad