Time Nick Message 00:25 bigfoot547 Hey, can I request a wiki account? 00:25 bigfoot547 Rather, may I have one? 00:26 jas_ are there two accounts for regular and dev wikis? or is it one for both i wonder 00:26 bigfoot547 we're about to find out :P 00:27 Aerozoic they're different 00:27 bigfoot547 Okay 00:27 bigfoot547 May I please have an account on both wikis? 00:50 jas_ i need to read the style guidelines, i presume those are on the dev wiki. would it not be easier if there was one wiki, and two sections between dev and user? 00:51 jas_ https://github.com/minetest/minetest_game/pull/1794/files#diff-a3d434548f6c2a40908f02f8a79a1c05R151 <-- palette_index*2^5 (normally seen in various source files) or palette_index * 2 ^ 5 (mmm spaces) 00:53 * octacian has returned from his rather extended break from MT 00:53 jas_ welcome back 00:54 octacian And by rather extended I mean 3 months rather than 2 00:54 octacian jas_: Thanks :D 00:55 octacian So, a question directed towards server owners: I'm adding a /ignore feature to my chat3 mod. Along with that, I am trying to provide a nice method to prevent players from ignoring specific players. Should non-ignorable players be registered via an API (i.e. chat3.prevent_ignore) or via a chatcommand? 00:56 octacian Shara, behalebabo: Would you be willing to provide some input on the above? 00:57 Aerozoic octacian, have you seen MG's ignore mod? 00:57 octacian Nope, link? 00:57 octacian Also, who' 00:57 jas_ https://forum.minetest.net/viewtopic.php?t=13993 00:57 octacian *who's MG? 00:57 Aerozoic octacian, https://github.com/MinetestForFun/ignore 00:57 octacian Ah, yes. Thanks. 00:58 rubenwardy server owners would likely want a privilege or setting 00:58 Aerozoic ah jas_ beat me to it XD 00:58 behalebabo octacian: it should be linked to an 'unignorable' privilege I think 00:58 jas_ i think there's a CSM ignore perhaps. and i'm looking forward to mod channels (?) so the server can see and manage this, but not do the actual processing 00:58 rubenwardy so unignorable as said 00:58 rubenwardy or ignore_override 00:58 rubenwardy hmm 00:58 behalebabo that can be set from both lua and chat 00:58 octacian I suppose I privilege wouldn't be a bad idea... 00:59 octacian Maybe I'll go for that 00:59 jas_ i can't help but think of the `shout' priv 00:59 rubenwardy it's a slight grey area, but I reckon it's fine 00:59 rubenwardy double_shout 00:59 jas_ double_plus_good_shout 00:59 rubenwardy yell 00:59 jas_ ahhhhhhhhhhhhhhh 01:00 jas_ joking aside, am i understanding this right? mod channels will allow server to manage/manipulate the client-side mods? 01:00 rubenwardy it's for sending messahes 01:01 octacian hmm. Should PMs be ignored as well? I'm doing that already now in chat3, but noticed it wasn't done in the linked ignore mod. 01:01 jas_ hrm. i'm interested in CSM but not if i can't send from server and then see/manage/manipulate sent CSM mods from server side 01:01 rubenwardy yes, octacian 01:01 jas_ and /me 01:01 octacian jas_: Yeah, I'm starting to think that CSM should have been implemented like that from the start, followed by implemented client-loaded CSMs. 01:02 rubenwardy you can write mods as if they're server sent 01:02 jas_ ok. i'm not concerned with whatever csm the client may have on their own behalf 01:02 jas_ i just want to send them ones specific to the server/sub-game, and then utilize them. like offload chat processing 01:02 jas_ /ignore is a good example, methinks 01:02 rubenwardy octacian: probably security related 01:02 rubenwardy urrr 01:02 octacian But, has there been any progress on server-injected CSMs yet? 01:02 rubenwardy octacian: probably security issues 01:02 octacian Like? 01:02 rubenwardy remote code execution, for one 01:03 rubenwardy that's literally what server injected CSMs are 01:03 jas_ oh yeah no i don't want that haha 01:03 rubenwardy so you need to make sure the sandbox is tight 01:03 rubenwardy and also limited 01:03 rubenwardy in resources 01:03 octacian I don't understand as to why that would be an issue, CSMs cannot make any physical changes to the client computer besides using modstorage. Last I checked anyways. 01:03 jas_ i just want to tell client, "ok execute ignore() on your own behalf, i'll just keep track of the fact that you did so." 01:04 rubenwardy our lua sandboxing may have holes 01:04 rubenwardy for example, if Lua has vulnerabilities 01:04 octacian Aside from that, the only thing that CSMs can do ATM is have a very minimal effect on gameplay. 01:04 octacian That's something where IMO you're better off just taking the risk and getting it done. Plus, if anything showed up it could be ironed out before release. 01:05 jas_ let's get this asio shit over with first, i feel like 01:05 rubenwardy it's most likely that it would only show up when exploited 01:05 octacian asio? 01:05 jas_ https://github.com/minetest/minetest/pull/6485 01:05 octacian rubenwardy: Exactly. So the couple months remaining until next release (or whenever it's planned) would give lots of time for people to exploit the bugs so they could be fixed :P 01:06 octacian jas_: True, that looks important, but no reason as to why it should prevent other things from being done :P 01:06 jas_ nerzhul does both this and csm 01:06 jas_ (i believe, and so may be either or) 01:07 octacian There's others who would likely be willing to work on some CSM features. It's not like CSM is a nezhul only thing :P 01:07 jas_ no doubt, i know but i'm talking about mod channels and csm ... what did you call it injection? i think that's his forte 01:07 jas_ i'd love it if more interested parties got involved, of course 01:08 red-001 I really need to fix my CSM prs 01:08 octacian rubenwardy: BTW, is there anything wrong with the possibility of ending up with 3 incompatible releases? 01:08 jas_ it's bad enough with two! 01:08 octacian You guys are way to worried about compatibility :P IMO it really holds back development speed 01:08 rubenwardy fragments the online community 01:09 jas_ i feel the opposite. slow and steady has served minetest well, i think 01:09 red-001 it's just really hard to fully secure anything 01:09 rubenwardy if it wasn't for Android, I wouldn't be too woried about it. As an online game, we're too relaxed on clients not being up to date 01:10 octacian It has only served MT well to some extent. Had this approach not be used, I personally feel as though we'd have progressed much faster in recent times. 01:10 red-001 even VM's get exploited 01:10 rubenwardy I think it would be better to download the mods from a third party 01:10 octacian Almost anything can be exploited if you try hard enough 01:10 jas_ that minetest is as accessible on as many different platforms/hardwares as possible is very, very important to its success, imho 01:10 jas_ and that all those versions kinda sort work together... 01:10 octacian rubenwardy: What do you mean, DL mods from a third party? 01:11 red-001 octacian, like a mod repo 01:11 octacian i.e. download mods directly from mod repo when connecting to server, or allow installing mods in general from the mod repo? 01:11 red-001 so you can check mods to see if they are doing any questionable stuff before uploading them to it 01:12 jas_ i'm confused are you guys talking about how players get mods for themselves to try out, or how servers send mods to clients? 01:12 octacian jas_: Incompatible versions doesn't affect the first thing you mentioned. The latter would be unimportant if we had some sort of launcher to handle version management. Although some different solution would have to be implemented for mobile users. 01:12 * jas_ misses the mod store 01:12 rubenwardy no, it'll be automatic 01:12 rubenwardy but from mods.minetest.net rather than the server 01:12 rubenwardy I started working on a mod store, but nerzhul disagreed with my tech 01:12 red-001 rubenwardy, so the server sends the mod id to the client and the client fetches it from the mods store? 01:13 octacian rubenwardy: So, you're saying download mods from a public repository when connecting to a server rather than from the server itself? 01:13 jas_ yeah because how does that work for a sub-game? and one that uses submodules with everything on github? that can't be what you meant... 01:15 rubenwardy client modding defines a global variable 01:15 rubenwardy so you can use the same repo for a SSM and a CSM 01:15 rubenwardy by checking whether it's running on the server or client 01:16 jas_ oh ok 01:16 octacian huh, I think I understand... 01:16 jas_ you're too honest 01:16 jas_ just smile and nod like i do 01:17 octacian End goal being to ensure that a player can read any CSM code that is to be injected? 01:17 red-001 anyway I have a feeling I'm pretty out of the loop here, I see the network system is being replaced? 01:17 octacian jas_: Nope, I wanna know for sure :P 01:17 rubenwardy is that guy still doing the mod store? 01:17 octacian red-001: You? Out of the loop? I've been away for 3months :P 01:18 red-001 octacian, look at my github account 01:18 red-001 enough said 01:18 octacian rubenwardy: There hasn't been much progress recently that I've seen. You think Python + Django would've turned out to be a suitable platform? 01:18 octacian I see.. 01:18 rubenwardy more than pure PHP 01:23 rubenwardy this actually needs to be done now 01:23 jas_ what does? 01:24 rubenwardy we should have a discussion on the models, the technologies, and the practicalities with celeron55 involved 01:24 red-001 if the network system is being replaced it might be a good idea to allow for server to server communication 01:24 rubenwardy but not at 02:30am 01:24 rubenwardy that's not really essential ATM, probs best to prioritise other things 01:24 octacian rubenwardy: Seems the CSS of your modding book is a bit messed up... https://i.endev.xyz/IMW9c5yf2P.png 01:25 rubenwardy that's correct 01:25 rubenwardy I centered it 01:25 rubenwardy does it look odd to you? 01:25 octacian Yeah. I personally preferred the old look 01:25 rubenwardy :( 01:25 octacian I dunno, it just doesn't feel right... 01:26 octacian Plus, the old view allowed for more to be seen on one page, which is definitely an advantage :P 01:26 rubenwardy it's easier to read narrower pages though 01:26 octacian Don't agree personally, but 01:28 rubenwardy I could left align it again 01:28 rubenwardy I may have been inspired by read the docs when originally making that style http://docs.readthedocs.io/en/latest/yaml-config.html 01:29 octacian I'd say either re-left-align it, or remove the background from the navigation, center it in the space remaining, and fix it to the page 01:55 jas_ server to server may not be essential, but it's still a very, very cool thing to do. i'm imagining with CSM and SSM fusion, it would be easy to minetest.disconnect(player) and minetest.connect(player, "coolerserver.net") 01:56 jas_ step into this portal, go to the creative realm. go to the survival realm. etc 02:03 bigfoot547 sooo... May I please have the accounts? 02:10 jas_ i believe calinou is the one to ask 02:10 jas_ i'd give them to you, but i don't have them. 02:13 rubenwardy Decided to stress test chunk sending in my game (Even though it's a placeholder). Fixer will enjoy these graphs: https://i.rubenwardy.com/5m37fc5gP4.png 02:13 jas_ i still think it'd be cool to have #minetest-offtopic and #minetest-docs; i know there's already some other channels, some unused. 02:14 jas_ maybe not -docs, haha 02:20 bigfoot547 I'm keeping -docs registered :P 02:20 bigfoot547 It seems like a good idea to have a docs team 02:21 bigfoot547 IMO 02:21 bigfoot547 I hope I share that idea with others LO 02:21 bigfoot547 :P* 02:44 octacian rubenwardy: Link to your new game? 06:26 nerzhul hi 06:26 nerzhul note for testers: i bundled asio in https://github.com/minetest/minetest/pull/6485 it's not recommended to use not up-to-date system asio library (1.10.6 and <) 08:13 jas_ ty. will test again 08:16 Calinou bigfoot547: send me a PM with your e-mail address and desired username (note that the username will start with an uppercase character, due to MediaWiki limitations) 08:16 Calinou (/msg Calinou ) 09:07 jas_ sweet, it works. 09:09 jas_ i'm going to put it on dcbl.duckdns.org, and already a client needs 0.5.0-dev; and now they will need 0.5.0-dev-asio haha! 09:17 Calinou hosting a sever on an incompatible dev version is asking for trouble :P 09:17 Calinou server* 09:21 jas_ it's for my own purposes, so it should be ok 09:21 jas_ what do you think? i'm not expecting any more players than i normally have 09:22 jas_ my pinky finger is cold from frosting my superheated pc 09:22 * jas_ heads for the lukewarm water 09:43 nerzhul jas_, you are crazy :D 09:44 nerzhul but it's nice if you use my asio patch we need testers 09:44 nerzhul but nobody can connect on your server :p 09:44 jas_ i can : ) 09:44 nerzhul my server is also with asio but not in public list 09:44 jas_ i will test it soon 09:44 nerzhul if you can run it in debug mode 09:44 jas_ yes that's what i'm working on now 09:45 jas_ getting my gdb args file set logging learning on 09:45 nerzhul :) 09:45 jas_ still don't know this program at all. and i guess --terminal is out with it? 09:45 nerzhul i run my server in tmux 09:45 nerzhul gdb ../bin/minetestserver 09:45 nerzhul > r 09:45 nerzhul :p 09:45 jas_ i like issuing commands to the server externally from terminal 09:45 jas_ oh 09:45 jas_ just > file huh? 09:48 jas_ nerzhul: i thought you linked your server on that PR comments page, but i don't see it now. it's just your server:30001 ? 09:49 jas_ found it. https://i.imgur.com/pSxrrbd.png 09:49 jas_ feels like i crashed ur server 09:50 Krock RIP 11.73111 hours of uptime 09:50 jas_ now i've reconnected A-OK but it's _all_ grey 09:50 * Krock wonders why jas_ uses this hardly readable font 09:50 jas_ uptime 42337 09:50 jas_ i was wondering when you'd ask 09:51 jas_ j/k 09:51 jas_ it only has ascii 09:51 Krock even worse 09:51 jas_ <3 09:51 Krock use comic sans 09:51 jas_ i love it when you say that HAHA 09:51 Krock features about the same set of characters 09:51 jas_ i pick on ppl who use that, so it's funny 09:52 jas_ anyway, i like this atari font i found in deb repo 09:52 jas_ i'ma keep it for a bit 09:53 Krock When you patch a bug but the feature still doesn't work as it should for two times: https://i.imgur.com/JKjR67P.jpg 09:56 jas_ oh i get it 09:56 jas_ memes just got the text removed from the picture 09:56 jas_ so now you just say it first. and so when there's text of a meme, with text, that's meta? 09:56 jas_ yea i'm definitely hanging on "Item definitions..." 10:06 nerzhul jas_, it's possible heh :) 10:07 nerzhul jas_, if the server has crashed it's "normal" to hang on this item defs 10:07 nerzhul in fact i need to find the proper way to know TCP connection is established, not only socket opened, it's not simple with ASIO by default 10:07 jas_ i'll keep trying 10:07 nerzhul my server doesn't crashed 10:08 jas_ oh btw, what's the preferred bt command to issue ? 10:08 jas_ i mean locally 10:08 sfan5 thready apply all bt full 10:08 jas_ thanks again, i'll post-it to my forehead 10:08 nerzhul i just do "bt full" it's generally on the right thread 10:08 nerzhul but having all thread in asio context is important 10:10 nerzhul i restart my server, it's not this morning version (which only bundle asio) 10:11 nerzhul sfan5, did you every see a such link error on mingw ? https://travis-ci.org/minetest/minetest/jobs/283070061 10:11 sfan5 sounds like you need to link to windows socket lib 10:12 nerzhul asio should use more high level windows API than we used before then ? 10:13 sfan5 idk 10:13 sfan5 it seems to be using some extended stuff mingw does not provide / doesn't support correctly 10:16 nerzhul wsock32 is seems somes doesn't link this lib 10:17 Krock jas_, indeed, that could be considered as background or meta information 10:19 jas_ cool, i've always sucked at memes, but always enjoyed them. mm cat pics. 10:22 tenplus1 hi folks 10:25 jas_ hiya ten 10:25 tenplus1 hi jas 10:25 jas_ how are things? 10:25 tenplus1 busy... :P you ? 10:26 Krock hi tenplus1 10:26 jas_ ohh, not too busy. relaxing to a cup of peanut butter banana coffee 10:26 tenplus1 hi krock 10:26 tenplus1 ooh, that sounds nice jas 10:26 Krock jas_, your interest ---> https://fireglow.de/755/kittens/ 10:26 tenplus1 Mobs Redo API has been updated so that flying/swimming mobs no longer get stuck in stairs/plants 10:26 jas_ it's so gross. my regular coffee is gwooonn, and my coworker gave my this elvis coffee a year ago 10:26 jas_ *cough* 10:26 jas_ ok do they still swim atop water? 10:27 jas_ thanks krock! 10:27 tenplus1 if a shark attacks it can FLOP out of the water and become useless, untill you punch it back into the water again :D 10:27 jas_ haha yeaaaa i'm saving this for caturday 10:27 jas_ i are serious cat. this is serious thread. hahahaa 10:28 tenplus1 heh 10:28 jas_ no i mean, do the cows still walk on water? 10:28 Krock sharks can do Floating point Operations? 10:28 jas_ or is that just me 10:28 tenplus1 oh, cows n sheep can float on water yes... but water hurts them slowly 10:29 jas_ i'm planning on making them sink somehow, why don't they out of curiousity 10:29 jas_ the zombies sink i think 10:29 tenplus1 they dont sink because sheep and cows can swim irl 10:30 jas_ i use opaque water and it looks sooo funny 10:30 shivajiva Hi folks :) 10:30 tenplus1 hi shiva:) 10:30 jas_ hi shivajiva 10:31 tenplus1 if a bat flies into water now it gets stuck and sinks :D 10:31 tenplus1 hi Jordach 10:32 shivajiva tenplus1: added sauth to my github account, have a peek sometime 10:32 jas_ it's weird i've announced to servers.minetest.net with asio patched minetestserver, as it states in action log -- but i don't see it listed on servers.minetest.net 10:32 tenplus1 kewl shiva 10:33 tenplus1 fixed mob api so they dont get stuck on stairs/plants :D 10:33 shivajiva :) 10:33 shivajiva that will make it harder lol 10:33 jas_ ew, minetestserver was racing my cpu 10:34 jas_ so, just to confirm nerzhul, one should NOT use libasio > 1.10.6 ? 10:36 jas_ also, in CMakeCache.txt, the comment above ASIO line reads: "// Enable using a system-wide JSONCPP...." 10:37 shivajiva anyway to tell if an on_put was a right click inside the event on a detached inventory? 10:38 jas_ think that's client-side 10:39 shivajiva it's making sorting an inventory contents practically impossible 10:40 jas_ what is it you want to do? 10:41 shivajiva 8 slot inventory, place stack in correct slot, handling partial stack and right click in the wrong slot successfully 10:42 jas_ so maybe some CSM improvements can be made, such that left-click/right-click/shift-click/drag-click from client can be handled from server-side mod. 10:43 jas_ or, observed and monitored, or something. 10:43 shivajiva it's the right click event that's eluding me, very difficult to prevent duplication in that scenario 10:44 jas_ that information is unsent to the server, afaik 10:44 shivajiva yup that's how it feels when trying to handle it 10:46 tenplus1 hi Dar 10:48 shivajiva will have to be satisfied with only allowing placement in the correct slot and handling the stack drop 10:49 jas_ i _believe_ the leftover stack that they're holding is actually still represented in their inventory list, and you should see its get_count() reduce when one right-clicks off from it 10:50 jas_ but a CSM<=>SSM bridge sounds cool too 11:00 nerzhul CSM <=> SSM bridge is merged :p 11:00 nerzhul it's mod channels 11:01 nerzhul jas_, asio 1.10.6 works but has warnings, 1.10.8 is the up to date versoin and is now embedded in lib/ if you want to use system lib add -DENABLE_SYSTEM_ASIO 11:01 jas_ oh good, ok. i know nothing about it yet, time to find it. sorry for my ignorance! 11:01 jas_ ok, am going to use the bundled version. my repo version is 1.10.8 also, but i've just removed it 11:01 nerzhul mod channels permits CSM and SSM mods to join communication channels to exchange data between paired client server 11:02 jas_ saweeet, shivajiva we just need a csm to read right-clicks :D 11:02 nerzhul you can have then CSM to CSM communication between clients 11:02 jas_ private messages! 11:02 jas_ i think this is really exciting. bravo! 11:02 nerzhul jas_, for example, one client can for example detect something arround him and share with other users 11:02 jas_ all without burdoning the server, which is fantastic 11:02 nerzhul sfan5, https://travis-ci.org/minetest/minetest/builds/283129711?utm_source=github_status&utm_medium=notification nice windows build is fixed 11:04 nerzhul it was the missing wsock32 lib :p 11:04 jas_ nerzhul: the line in CMakeCache.txt reads // Enable use of system-wide JSON-CPP..." for ASIO option FYI 11:05 nerzhul oops :p 11:05 jas_ : ) 11:14 tenplus1 hi fixer 11:23 Jordach nerzhul, what's SSM? 11:31 shivajiva serverside mod 11:34 tenplus1 how about mmM 11:36 Fixer ohi 11:37 tenplus1 hi Megaf 11:37 Megaf Heya 11:37 Megaf Got my Raspberry Pi Zero yesterday 11:37 Megaf talking from it right now 11:40 shivajiva or not ^ 11:40 tenplus1 ahaha 11:41 jas_ no joy for me with asio, yet. 11:41 tenplus1 I see many videos on youtube for pi3, especially the overclocking and cooling 11:42 jas_ can't connect to unix-experience.fr:30001 (all grey, server unresponsive); can't run minetestserver here because it hangs and sends my CPU to 100% 11:42 tenplus1 eek 11:46 tenplus1 hi thomas and Aerozoic 11:47 Aerozoic wazup 11:47 ThomasMonroe hi tenplus1 11:47 tenplus1 o/ 11:51 jas_ next up to test: mod channels 11:51 * jas_ wraps fingers 11:52 tenplus1 back laters o. 11:56 Megaf lol, dont do /sysinfo on hexchat running on a raspberry pi 11:56 jas_ but i don't have those things! 11:57 Megaf I'm talking from a Raspberry Pi zero 11:57 jas_ oh yea? 11:57 jas_ how's that? 11:57 Megaf cant believe I paid 4 pounds for a computer 11:58 Megaf 4 pounds and it runs LibreOffice, Chromium, Hexchat, Firefox, GIMP, Inkscape, heck, even blender 11:59 nerzhul jas_, all grey ? 11:59 nerzhul yes the server CPU problem was reported in my main task list, i will look at it this evening 11:59 nerzhul for the connection i see you connect and disconnect fast 11:59 jas_ ok 11:59 jas_ yes, i am presented with grey in front of me, and 1 mapblock's worth of desert behind me 11:59 jas_ i can dig and jump on it, but no /status 12:00 Megaf So, the Pi zero is quite nice, you just have be sure to use a decent sd card with it 12:00 Megaf otherwise you gonna be stuck in lots of IO wait 12:00 jas_ slow huh? 12:01 Megaf Its 1 GHz single core, but it actually runs the OS pretty well 12:01 Aerozoic Megaf, where'd you find it so cheap? 12:01 Megaf they optimzed Debian for it 12:01 Megaf pimoroni 12:01 Megaf I couldnt believe myself, they had in stock and it was 4 pounds 12:02 Aerozoic what's that? a local store? 12:02 jas_ unfathomable 12:02 Megaf 3.33 pounds shipping and 1.,47 VAT 12:02 Megaf pimoroni.com 12:02 Aerozoic oh ok 12:02 Megaf brb later 12:03 jas_ and i didn't know how to show that, either nerzhul. i think i took a screenshot 12:03 jas_ i mean, i tried bt command, but of course it was no error 12:03 jas_ server just didn't get anything. maybe i should run minetest with --trace ? 12:05 nerzhul there is not many logs, only gdb breakpoints are generaly useful 12:05 jas_ ok 12:05 nerzhul i will look at the problem this evening, i think either server or client is in a strange stage, maybe i forgot something :) 12:06 jas_ great, we'll try again then 12:06 jas_ now i'm off to learn about csm and channels 12:31 lisac nerzhul, do you think MT should maybe be in Hacktoberfest? https://hacktoberfest.digitalocean.com/#details 12:31 lisac seems like a nice way to attract contributors. 12:54 Fixer depends on their impression 12:55 Fixer just don't show them mtg to deep :trollface: 12:55 Fixer too* 12:55 Fixer but from a distance 12:56 Fixer MTG: 3 merged PRs in 1 month 12:57 Megaf lovely 12:57 Fixer paramat is top contributor as usual 12:58 Fixer one of the most active OSS projects: Cataclysm DDA 12:58 Fixer 154 merged PRs in 1 month 12:59 Fixer minetest engine: 31 merged PR in 1 month 12:59 Fixer not bad 12:59 Fixer but MTG is just halfdead 13:00 Fixer it needs more contributors 13:00 Fixer current list is not enough 13:00 Fixer maybe some modders are interested 13:00 ThomasMonroe Caleb would be really interested 13:01 CalebDavis once i get a GH account XD 13:01 ThomasMonroe XD 13:01 ThomasMonroe you canuse mine for the time being 13:02 ThomasMonroe just make sure you add a -Caleb when you use it XD 13:05 IhrFussel Hi guys 13:05 ThomasMonroe hey 13:06 CalebDavis hi Ihr 13:07 nerzhul lisac, i don't care 13:08 Krock does anyone know if switching from WTFPL to CC0 is possible in the regular cases? 13:08 Krock Like, when other people already contributed under the WTFPL license 13:09 red-001 yes 13:09 red-001 WTFPL has no limitations 13:10 red-001 if you wanted to you could even make the code closed source 13:14 Krock well, the limit is the warranty disclaimer, which is handled in CC0 13:15 nerzhul WTFPL = do what you want, then you can make it non-free if you want 13:15 Krock hehe.. *changes it to -NC * 13:27 jas_ ok so we have client-side, channels, and server-side. when can we send mods from server to client? is that the big security concern? 13:28 jas_ can the server direct the client to (optionally) get the client-side mod from a specified URL? 13:29 jas_ in quake(sorry paramat), one would optionally download from the server. it was off by default, but once you turned it on you could download the mods from the server 13:29 jas_ sv_autodownload or something, calinou would know LOL 13:33 jas_ o.m.g. i'm typing quake command cl_, sv_, into GOOGLE, and it's listing them in realtime hahahaha 13:35 jas_ sv_autodownloads 1, that's it. so yeah, would just putting a setting, default off, like download_mods_from_server or something :D 13:36 jas_ ? 13:39 sfan5 csm is sandboxed, so servers sending lua code should not be a problem 13:40 jas_ nice. well i'm stoked, i printed off a fress lua_api, that channels png, and client mod doc raw. just need the client to retrieve a mod from server automagically 13:41 Calinou Krock: if you follow the WTFPL to the letter, yes, you can 13:42 Calinou jas_: client-side mods can't be sent from server to client, that would be too insecure 13:42 jas_ you have to follow local ordinances in your jurisdiction firstly, though. 13:42 Calinou id Tech 3 games have many known security vulnerabilities with this (the QVM isn't 100% secure) 13:42 jas_ jfc calinou 13:42 Calinou (it was even worse with id Tech 2, where you just get native .dll/.so files) 13:42 jas_ somebody call me a waahhmbulance 13:42 Calinou jfc? 13:42 Krock okay, thanks nerzhul and Calinou 13:42 jas_ so what is a hopeful dreamer like me to do then?? 13:43 jas_ say to anyone who connects, "hey can you go to this website and download this zip file and extract it into this folder?" jeeeeeeeeeeeez 13:43 * jas_ cries and goes off to pace or something hehehe 13:53 nerzhul Krock, for ? 13:54 Krock license answer 13:54 nerzhul jas_, currently i'm working on the network, when ready and (hope) merged i will look at the CSM mod sending and propose a draft concept 13:55 jas_ i'm one click from creating an issue, i just wrote. 14:04 sfan5 sending client side mods isn't even implemented yet so that issue is not very useful 14:06 nerzhul yes and there is an issue about that :) 14:08 jas_ oh the setting part? 14:08 jas_ no, yeah, i didn't know there was that issue already 14:08 jas_ sorry 14:09 Krock aww shit. I missed my 666th contribution this year 14:13 Krock jas_, now send this mod code to a client to execute: while true do end 14:13 jas_ ok thanks for the lively conversations, sorry for my confusion again. i think i can just check if the player has the right CSM and kick if not, for now 14:13 Krock they'll be quite busy 14:14 jas_ it's not my problem 14:14 jas_ i don't do that... anymore HAHA 14:15 jas_ i mean, accidents happen. but i wouldn't expect a connecting player to execute i wouldn't execute 14:16 jas_ so if it's tested, and frankly identical to what i use, i can think the other connecting clients would be safe too 14:16 jas_ because with this i can offload stuff from server, and do the work on the client side. 14:17 jas_ it'll be like splitting out a lot of dcbl, and having the server expect the client to have the dcbl mod, or whatever. and it won't work if it doesn't match 14:17 Krock or let's implement a bitcoin farm by Lua injection. Android players won't notice the additional lag :P 14:18 jas_ it's fine if you're opposed to sending mods. don't do it then. turn the setting off by default. 14:18 jas_ or, i can kick any connecting player who lacks this CSM my server needs them to have, with a message where to get it 14:18 Krock I'm just trying to point of the possible abuse by such an opportunity 14:18 jas_ i think we can compromise 14:18 jas_ i see your concerns, but they're not mine 14:19 jas_ i think a dialog in builtin, with a security message warning, that would ONLY appear in the first place IF csm_autodownloads was true, stating, "ARE YOU REALLY FOR CERTAIN?" yes or no 14:19 jas_ then, it's really on them? i don't know, i hope i'm on track here 14:24 jas_ [Agree] [View Code] [Cancel] 14:29 red-001 if you want something like that implement it in C++ not Lua 14:29 jas_ or not view b/c you'd probably end up parsing it once gotten 14:30 jas_ unless in C++ it would be safe 14:31 red-001 still I think that's pointless 14:31 jas_ viewing? 14:31 red-001 people will just ignore the warning 14:31 jas_ oh the warning. well it relieves burdon, that's not for nothing 14:31 jas_ a warning shifts liability 14:31 jas_ you'd prefer not being able to send CSM to client? 14:32 sfan5 people will just routinely click security warnings away 14:32 sfan5 this achieves nothing 14:32 red-001 I'm saying if that's the security you want to use you might as well not bother 14:32 jas_ ok, i'll drop it then and wait and see. 14:32 sfan5 "shifting liability" is also pointless unless you plan on going to court about client-side mods 14:33 jas_ "security" 14:33 Krock relevant: http://pointcost.com/technology/client-side-security/ 14:34 nerzhul Krock, i suggest ethereum :p 14:34 nerzhul hey red-001 14:35 red-001 hi nerzhul 14:36 red-001 if we secure the sandbox well the biggest risk will be someone finding a bug that can make Lua generate unsafe bytecode 14:38 nerzhul red-001, where are we with your CSM patch ? do we merged all security concerns and features ? If i remember hud is not merged, red-001 can you update https://github.com/minetest/minetest/pull/6067 ? 14:38 nerzhul and https://github.com/minetest/minetest/pull/6087 14:38 nerzhul it's important to have them asap :) 14:42 jas_ orwell96 was asking about dofile 14:42 jas_ oh nvm im blind 15:01 red-001 wtf why is singleplayer refusing to let me connect 15:02 red-001 so strict version checking is broken it seems 15:03 Krock not sure if that's a funny or weird bug :3 15:03 red-001 I would say both 15:04 red-001 it told me to contact the server admin so I contacted myself 15:05 Krock any success? 15:05 Krock I did the same too but that lazy dude never responded or cared about it 15:06 red-001 well the admin though to check if strict version checking is on and just disabled it 15:06 red-001 and marked the issue as fixed 15:06 Krock he's not wrong 15:08 red-001 he seems to have a real cargo cult mentality towards the fix and refuses to investigate it 15:11 red-001 nerzhul, updated https://github.com/minetest/minetest/pull/6087 16:52 IhrFussel What's the default time a server waits before disconnecting a player on timeout? 30s? 16:53 IhrFussel And does the "player still there" check happen in intervals? Or does the client just say "hey wait I'm still here" whenever the connection is back? 16:53 nerzhul 30s if i remember 16:53 nerzhul it's more difficult in UDP but it's 30s or 1 min 16:54 nerzhul udp socket timeout is set to 5 sec on connection.cpp it seems at connection creation 16:56 IhrFussel 5s? But the server doesn't kick anyone after that time correct? 16:57 nerzhul it seems no it's a little bit complicated, the timeout code is not easy to understand in current model 16:57 IhrFussel ASIO won't change that I guess? 16:57 nerzhul tcp timeout, exactly 30sec 16:57 nerzhul after that socket close and disconnect user 16:58 nerzhul it's not perfect currently but it's the goal, when asio socket resource is closed, connection object is destroyed, serverconnection thread unregister client and tell server thread he disconnected 16:58 IhrFussel So when the client loses connection does it ping the server regularly trying to reach it and prevent getting disconnected? 16:59 nerzhul if he looses the connection, he looses the connection 16:59 nerzhul 30sec grace time 16:59 IhrFussel I mean when the connection is gone for a few secs...what does the client do in that time? 17:00 nerzhul like now, no event is received by server, and if keepalived is up before server timeout (TCP side or MT UDP current side) session is kept and should be fine 17:00 IhrFussel I'm guessing it's calling for the server and hopes to be heard xP 17:01 IhrFussel And if I saw correctly as soon as the connection is back (before the server closed the socket) the server seems to send all missed packets to the client...could that be right? 17:02 nerzhul with TCP it's right 17:02 nerzhul with UDP it depend on the implementation 17:02 nerzhul i think it's the case with MD protocol 17:03 IhrFussel Does MT have any priorities on the network protocol used? Like try TCP if fail try UDP if fail ??? 17:04 red-001 IhrFussel, right now it's all UDP 17:04 red-001 nerzhul is rewriting it to use TCP for relaiable data and UDP for input 17:05 nerzhul it's a poc 17:05 nerzhul it's working as i intend currently but not perfect 17:06 IhrFussel So MT only used UDP until now? Weird when I made a server people always told me to always choose both modes 17:07 sfan5 then they were wrong 17:07 sfan5 mt has always just used udp 17:10 IhrFussel So TCP is slower I just read 17:10 IhrFussel But it *guarantees* the successful of the transmission 17:10 nerzhul if you use it for game data which need speed and doesn't need reliability yes 17:10 IhrFussel success* 17:11 nerzhul for example FPS needs speed only, because you do only inputs (movement + clic) 17:11 nerzhul and loose 1 packet is not important 17:12 nerzhul IhrFussel, can you test my current PR with my server ? i need a remote feedback from germany 17:14 IhrFussel nerzhul, if you explain to me how to apply a PR that didn't get merged yet sure...I only know how to compile master 17:14 nerzhul go to mt source root folder 17:14 nerzhul wget https://github.com/minetest/minetest/pull/6485.patch && git am -3 6485.patch 17:14 nerzhul compile 17:14 nerzhul :p 17:15 Krock or git remote add nerzhul https://github.com/nerzhul/minetest.git; git fetch nerzhul network_asio; git checkout nerzhul/network_asio 17:16 nerzhul more compliocated but i will get my own origin :D 17:16 Krock however, easier to update. 17:18 nerzhul y 17:19 Fixer Krock: i want to test this too 17:19 Fixer leeeeeeeeeeeeeeeeets heear it 17:20 Krock ack. 17:22 IhrFussel Compiling, will take a few minutes but at least this is not my old Celeron anymore (took 45 minutes) 17:23 Fixer lol 17:23 Fixer rare 17:23 Fixer and i thought my 6-7 min full compile is slow 17:24 Krock heck many unused variable warnings 17:24 IhrFussel My dedi compiles minetestserver in under 2 minutes 17:24 Krock IhrFussel, from scratch? if so, nice specs 17:25 IhrFussel Yeah I always remove everything old when I upgrade ... the specs aren't that great I think..it's an AMD Opteron from 2010 17:26 IhrFussel I guess it helps a lot that there's no DE and only a few background processes 17:32 IhrFussel nerzhul, I see lots of "asio::" defined but not used 17:41 bigfoot547 Calinou: hey 17:41 bigfoot547 u here? 17:41 Krock nerzhul, works nicely already 17:42 IhrFussel I see no RTT in F5 ?? 17:45 IhrFussel Huh? RTT is gone in F5 or does it require the debug priv now? 17:46 nerzhul Krock, update and use embedded lib :p 17:46 nerzhul did you set ENABLE_SYSTEM_ASIO ? 17:46 Calinou bigfoot547: yes 17:46 Calinou bigfoot547: send me a PM with your e-mail address and desired username (note that the username will start with an uppercase character, due to MediaWiki limitations) 17:46 Calinou (/msg Calinou ) 17:46 Calinou :P 17:46 nerzhul ircSparky, rtt is gone atm 17:46 bigfoot547 ayy thanx Calinou 17:46 nerzhul i don't have the measurement atm i should find the correct way to have it 17:47 IhrFussel nerzhul, do I need ENABLE_SYSTEM_ASIO? I didn't set it 17:47 IhrFussel But I think it works else I couldn't connect to the server...I'm on it 17:48 IhrFussel So what's the important value here for testing? dtime_jitter? 17:52 Krock nerzhul, using default 17:52 IhrFussel terminate called after throwing an instance of 'std::bad_alloc' 17:52 IhrFussel what(): std::bad_alloc 17:53 Krock ASIO is not installed here 17:53 Krock previously I included the headers manually 17:53 IhrFussel Isn't ASIO included in the PR? 17:54 IhrFussel Ok so after bad_alloc I reconnected to your server nerzhul and all I see now is gray 17:54 red-001 it's a new lib thats needed 17:55 IhrFussel red-001, the lib is part of the PR I think 17:56 IhrFussel I didn't install ASIO anywhere but was still able to play on his server 17:57 IhrFussel But now I see just gray; no map, no HUD ... it's like the old bug on servers were it took several minutes for players to finally see the map 17:58 IhrFussel Oh so gray screen means server is down? 18:00 red-001 is anything being rendered? 18:01 IhrFussel Oh wait I was wrong the HUd gets rendered and hand too...and the chat does show the status and displays "IhrFussel joined the game" 18:02 IhrFussel I can also look around but the whole map is just a gray background 18:03 Krock IhrFussel, rejoin fixes it 18:03 Fixer c o m p i l e d 18:04 IhrFussel Krock, didn't fix it I rejoined twice now 18:04 Krock a e s t h e t i c c o m p i l e 18:04 IhrFussel Nope 3rd time still gray 18:05 Fixer nerzhul: here is my compile log of your asio thing, it has warnings https://pastebin.com/raw/2y09puxL 18:05 Fixer [20:46:17] did you set ENABLE_SYSTEM_ASIO ? 18:05 red-001 s i m p s o n w a v e 18:05 Fixer waaaaaat? 18:05 Fixer i need to set enable_system_asio? 18:05 Fixer fuck 18:05 sofar for those interested, there's a fantastic new puzzle box on "Inside the Box" by raven 18:05 Fixer ok 18:05 Fixer will visit later 18:05 sofar be prepared to be stuck for a while :) 18:06 IhrFussel Fixer, I don't think that's really mandatory...I was able to connect to his server just fine and walk around without that flag 18:06 Fixer i will recompile 18:07 IhrFussel I think ENABLE_SYSTEM_ASIO is only required if you want to use your own asio lib 18:07 Fixer IhrFussel: cores help 18:07 Fixer IhrFussel: more cores - faster compile 18:08 Fixer error 18:08 IhrFussel My dedi only has 4 cores @ 2.5 Ghz 18:09 Fixer nerzhul: added -DENABLE_SYSTEM_ASIO=1 , got error on compile https://pastebin.com/raw/fV62cSuL 18:09 Fixer No such file or directory 18:09 Fixer hmmmmhmmhm 18:10 Fixer will do without asio thing 18:10 IhrFussel If we would've to use our own asio lib he wouldn't include it as lib 18:10 red-001 dammit copy and paste is broken again 18:10 red-001 how do you mess this up irrlicht 18:10 IhrFussel IT WORKS I see the map again 18:11 Krock good question. Ctrl+V Ctrl+A Backspace Ctrl+V solves it mostly 18:11 Krock like irrlicht stacks the clipboard for some reason 18:15 IhrFussel It definitely works without the compiler flag or else I couldn't even connect I think 18:16 red-001 getting multi second freezes 18:16 red-001 and it seems to drop connection randomly 18:19 Fixer where to connect? 18:19 IhrFussel My connection didn't drop yet, only had 1 segfault 18:20 IhrFussel Fixer, web.unix-experience.fr:30001 18:20 Fixer ty 18:21 Fixer lol 18:21 Fixer crassh 18:21 Fixer nerzhul: does not connect, instant CTD 18:22 Fixer no backtrace for ya 18:22 IhrFussel Why doesn't my client crash?? 18:22 red-001 IhrFussel, did you see any messages in chat? 18:23 red-001 I tried sending a few to see if it was working 18:23 nerzhul Fixer, i don't recommend using system asio if not 1.10.8 18:23 Fixer nerzhul: i don't have any 18:23 IhrFussel Yep I saw them now 18:24 nerzhul IhrFussel, i read changelog and see older asio version have problems, also asio 1.10.1 (ubuntu 14.04) is from 2013, 1.10.8 is 2016 18:24 nerzhul red-001, i see your messages in my console 18:25 red-001 ok I was wondering if it was working 18:25 IhrFussel nerzhul, but if we don't use the asio flag which version will MT use? 18:25 red-001 since I didn't see a good few of them myself 18:25 nerzhul 1.10.8 is in lib/ 18:25 IhrFussel That's the one I got then 18:25 nerzhul 1.11 is in heavy development because asio 1.11 is working to be reliable with C++20 draft for networking 18:26 nerzhul it will permits developpers to have smooth path to migrate to C++20 18:27 IhrFussel nerzhul, now what value do I use to test the network performance from Germany? dtime_jitter? 18:27 nerzhul IhrFussel, excellent question 18:28 nerzhul those indicators are currently removed because i trashed the whole stack, but i will re-add them when i fixed remaining points 18:28 nerzhul i think i can have bandwidth indicator but jitter... 18:29 nerzhul it's default mtg and mt server config IhrFussel :p 18:29 IhrFussel jitter is about the same compared to my 0.4.16 server if that helps 18:29 nerzhul Krock took 2 second to connect it's nice 18:30 nerzhul IhrFussel, nice 18:30 nerzhul and disconnection is instant too 18:30 nerzhul brb 18:30 Krock btw, nerzhul: this is the current map: https://i.imgur.com/5IjgTd0.png 18:31 Krock but I messed up the 2nd "B" 18:32 Calinou 151.25.0.0/16? 18:32 IhrFussel Disconnecting from Germany took 1 second here too (maybe even less), connecting takes 2-3 seconds 18:33 IhrFussel Did I disconnect? Map doesn't load anymore 18:34 Krock yes you did 18:34 IhrFussel Wow that reminds me of the days when even 0.4.X used to do that (disconnecting without timeout message clientside) 18:35 Fixer Krock: RESIST 18:35 Krock IhrFussel, join-quit again 18:36 Krock Fixer, fixed it https://i.imgur.com/3afjWbx.png 18:36 IhrFussel Yea I see exactly 1 mapblock in the distance 18:37 IhrFussel Happened yet again hmm 18:37 IhrFussel What the?? It happens every time now I even re-opened my client 18:39 Krock minetest 3.3 GiB ram. neato 18:40 Krock HDD was quite busy 18:44 Fixer SO DO I... 18:44 Krock ...WRITE IN CAPSLOCK 18:44 IhrFussel Now I only see gray again meh 18:45 Fixer waaat? 18:45 Fixer compiled debug build 18:45 Fixer connecting 18:45 Fixer says client disconnected... 18:45 * Fixer pokes nerzhul 18:47 Krock the server could be down? 18:47 Calinou Version 1.1.0 of More Blocks has been released! https://github.com/minetest-mods/moreblocks 18:47 Calinou includes a fully rewritten README :P 18:47 Krock Calinou, where's the forum topic - and more important? Where's the cake? :P 18:48 Calinou (feel free to use it as a template for your own mods) 18:48 Calinou I'll make a forum post on the More Blocks topic 18:49 Krock I got a sudden flashback while reading "Enable the mod". It's like Windows vs MacOS vs Linux: "Open start menu, then run.." vs "Go to the menu, step 1.." vs "Open the terminal." 18:50 sofar Calinou: nice, I see you're taking care of moreblocks :) 18:50 Calinou sofar: yeah, it needs some love :p 18:50 Krock indeed. it's a nice mod 18:51 Krock where are the dependency explanations btw 18:51 Calinou hmm, I forgot to add them, I can do that 18:51 Calinou also, it's been a long time since I didn't have a dark green square on my profile: https://github.com/Calinou 18:51 Calinou the last one was on June 16th 2017 18:51 Fixer https://github.com/minetest-mods/moreores/pull/27 18:53 Fixer hmmhmh 18:55 Fixer waaat 18:55 Fixer so debug build just says client disconnected and release one has CTD 18:56 Fixer nerzhul deleted a comment from kwolekr 2 hours ago 18:56 Fixer O_____O 18:58 Krock yay! the Windows drawing effect is back! https://i.imgur.com/ZJ8L39R.png 18:58 Krock inb4 swap use from 2% to 18% 18:58 Krock nerzhul, there could maybe be a memory leak while disconnecting from the server 19:00 sofar Shara: can you auto-voice jas_? 19:00 bigfoot547 ooo 19:00 jas_ thanks you 19:00 Shara Just doing it 19:00 sofar didn't realize I couldn't 19:01 bigfoot547 ono 19:01 bigfoot547 mutiny? 19:01 bigfoot547 *gasp* 19:01 bigfoot547 ChanServ: How dare you! 19:01 Shara Heh 19:01 Shara jas_: Can you try rejoining now please? 19:01 sofar atorian37: Flitzpiepe mransom Xio: PM an op to get voice if you need 19:02 Shara atorian37 is already on the list, but I guess didn't auth. 19:02 Shara Flitzpiepe! You managed to sneak in here without me noticing you! :) 19:02 Flitzpiepe I'm sorry Shara :) 19:03 Shara No problem, and welcome. 19:03 Flitzpiepe thank you very much :) 19:03 Shara Flitzpiepe has helped out on my servers, and on some others as well. 19:03 IhrFussel nerzhul's server is back on 19:04 bigfoot547 oooo 19:04 bigfoot547 Who's weqqr? 19:04 bigfoot547 Also, I have set server_announce true and done all the other stuff stated on https://www.minetest.net/servers/, but my server still won't announce 19:05 Krock wrong settings file, bigfoot547 19:06 bigfoot547 Which one do I put it in then? 19:06 Krock try changing this setting in world.mt 19:07 Krock unless you changed this setting over GUI.. then I have no idea what went wrong 19:07 IhrFussel I should probably compile a debug build...what's the compile flag again? 19:07 bigfoot547 Okay 19:08 bigfoot547 I see this line: 2017-10-04 14:08:33: ACTION[Server]: Announcing to servers.minetest.net 19:09 Krock IhrFussel, uhm.. -DDEBUG ? 19:09 bigfoot547 -CMAKE_BUILD_TYPE=Debug 19:09 bigfoot547 :P 19:10 bigfoot547 And on gcc/++: -g -DDEBUG 19:10 jas_ careful 1hr, watch your CPU 19:10 Krock bigfoot547, sounds good now. Perhaps tweak the server URL setting now (depending if you have a domain for it) 19:10 jas_ Shara: ok 19:11 jas_ oh great \o/ thanks shara 19:11 bigfoot547 Well 19:11 IhrFussel bigfoot547, it says file not found 19:11 bigfoot547 It's not on the server list 19:12 IhrFussel I mean "CMake Error: Error processing file: MAKE_BUILD_TYPE=Debug" 19:12 Shara You're welcome, jas_ 19:13 bigfoot547 Oh 19:13 bigfoot547 -DCMAKE_BUILD_TYPE=Debug 19:13 bigfoot547 Lol 19:14 jas_ i'm gonna try network_asio again. this time trying to connect to 127.0.0.1 minetestserver (i only tested connecting to remote servers, earliar) 19:14 IhrFussel Thanks^^ 19:14 jas_ earlier haha 19:27 IhrFussel jas_, his server seems to be down again 19:27 jas_ OKAY! asio works locally 19:27 jas_ doing minetestserver --world test, and minetest --go 127.0.0.1, we have joy 19:28 jas_ seems to use more CPU than regular networking? 19:28 IhrFussel I think local is what nerzhul worked on first 19:28 jas_ i'd imagine. surprised i didn't think to test earlier 19:28 jas_ but my true test is a remote connection (well, on my lan) 19:28 jas_ hrm, i'll try again 19:28 jas_ now that there was a commit made since i last tried 19:29 IhrFussel jas_, do you also get a gray screen + HUD if you try to connect to nerz's server? 19:30 jas_ earlier i did, yes. 19:30 Fixer https://pastebin.com/raw/UmzLFTKx 19:30 jas_ but for me, i wasn't even able to connect because when i launched the server, it would hang at 100 CPU 19:30 IhrFussel It worked now...I wonder if the server is really down then or if it's a display/network bug 19:30 Fixer nerzhul: https://pastebin.com/raw/UmzLFTKx when starting local game 19:30 jas_ no check the uptime 19:30 Fixer nerzhul: and connecting to your server 19:31 Krock nice paste, Fixer: https://i.imgur.com/yLCT8Mw.png 19:31 IhrFussel Oh yeah okay well then it's a random network bug to get the gray screen 19:32 IhrFussel *waiting for the next bad_alloc or segfault* 19:34 jas_ IhrFussel: start a minetestserver, and connect to it with a client on a different host! 19:34 jas_ that's what i'm going to try now. earlier, it would cause minetestserver to hang at 100 percent. dang i keep repeating myself haha 19:34 Calinou so, I was voting on the LibreOffice mascot proposals: https://survey.documentfoundation.org/665628 19:34 jas_ oh? 19:34 Calinou I ended up not voting anything… but one of them saves the day 19:34 Calinou https://media.hugo.pro/chrome_2017-10-04_21-29-15.png BEST WAIFU 19:34 Calinou (no, really… a lot of them are just too generic or poorly drawn) 19:34 Calinou sfan5: ^ 19:35 sfan5 >green 19:35 Krock ^ prefers pink hair 19:36 jas_ https://i.imgur.com/vGcpRsp.png 19:37 Krock upvoting the waifu now 19:37 Calinou https://media.hugo.pro/quasselclient_2017-10-04_21-37-11.png 19:37 Calinou jas_: same lines, but on my setup :P 19:37 jas_ i thought it was funny 1) the arrows, and 2) the pink 19:38 Krock coincidence 19:38 nerzhul Fixer, interesting 19:38 jas_ i love a good coinkydink 19:38 nerzhul jas_, i fixed the cpu problem 19:38 IhrFussel I was able to swim 500 nodes in one direction before it disconnected me 19:38 nerzhul before when no packet received, server waits for packet 30ms 19:38 jas_ good good, i'm compiling now 19:38 nerzhul 30s* 19:38 rubenwardy does the survey stop at somepoint? 19:38 nerzhul now i added a sleep time when no packet received, 50ms 19:39 Fixer nerzhul: windows 7 19:39 IhrFussel nerzhul, do you know what causes the random disconnects yet? The client never receives the message that it got disconnected everything just stops loading 19:39 nerzhul IhrFussel, to be exact in MT there is no local/remote network is used in singleplayer too 19:40 nerzhul IhrFussel, it's a point in my todo list i should reproduce 19:40 nerzhul server didn't crashed, it's always started i see, then you loose connection to server , or maybe there is a reading problem 19:40 rubenwardy it does 19:40 jas_ that's what i thought. doesn't make a difference whether minetestserver && minetest, or just minetest singleplayer mode; they're effectively the same (?!) 19:40 rubenwardy stop that is 19:40 nerzhul minetestserver is just minetest without UI & renderer :p 19:41 jas_ dunno why i got it into my head that the behaviors were different. oh well 19:41 nerzhul i think the stop loading problem is due to some weird unhandled tcp reading problem 19:42 nerzhul it's why remote test is important 19:42 octacian jas_: I just use minetest --server on my servers, far easier to keep up-to-date than when dealing with the minetestserver package IMO. 19:42 IhrFussel nerzhul, WHAT THE?! I just reconnected and now I get "You died"? 19:43 jas_ octacian: i have three computers with different purposes. one prod, one dev, and one laptop to interface and use regularly 19:43 jas_ the laptop has both minetestserver and minetest, tho here i am mostly just using the client 19:43 jas_ sometime i do like to test minetestserver, but i know i can just minetest --server 19:43 jas_ on the prod and dev boxes, it's only minetestserver 19:43 IhrFussel Did the server still keep track of my health even though I was supposedly disconnected? 19:44 jas_ but it is funny that i bother to build a minetestserver on this laptop when it's identical to minetest --server LOL 19:44 jas_ guess i can stop doing that. good point octacian 19:44 Krock IhrFussel, it won't sink.. so you don't have any further damage. the previous HP will be loaded 19:44 octacian Yeah, I just build the plain minetest package on all my computers, handles both singleplayer, client, and server rather than having to deal with multiple :P 19:44 Krock o/ octacian 19:45 octacian Hey Krock :D 19:45 jas_ i rebuild often, so on dev and prod, only recompiling minetestserver is way quicker 19:45 jas_ not building minetestserver on this laptop will save me some time too 19:45 IhrFussel Krock, but I think that's what happened...I likely disconnected underwater but with almost full HP and when I rejoined I was dead 19:45 octacian Usually make only compiles unmodified files, so build times between minetest and minetestserver should only differ very minimally 19:46 octacian Assuming that you are re-building (updating), that is 19:46 sfan5 octacian: minetest and minetestserver are built separately because their build flags differ 19:46 sfan5 so it will compile most of the code twice if you build minetest *and* minetestserver 19:46 jas_ i connected, but it's slow? i see krock and ihr are there 19:47 jas_ oh, i can't send chat? 19:47 IhrFussel jas_, you disconnected right after joining 19:47 octacian sfan5: huh, Good point. I forgot the flag for minetestserver, what is it again? 19:47 sfan5 -DBUILD_SERVER=1 19:47 jas_ it's because i have a new player name? 19:47 octacian Alright, cause come to think of it, for my console servers, it's pointless to build the full thing. 19:49 IhrFussel jas_, Nope it happens randomly 19:49 jas_ interesting 19:49 jas_ well the server is really shaping up anyway, i can tell it'l be popular 19:50 nerzhul IhrFussel, if tcp socket is disconnected you are instantly disconnected 19:50 sofar ASIO is tcp, right? 19:50 nerzhul but if there is a reading problem server side, (nothing loading) i think it's a socket reading problem, i should verify the reading implementation to see if sometimes there are side effects 19:50 nerzhul sofar, no 19:50 nerzhul asio is what you want 19:51 sofar ah 19:51 nerzhul asio = asynchronous IO 19:51 nerzhul you have tcp::socket, udp::socket 19:51 nerzhul and you can implement asynchronous wtf on what yo uwant if you need for example asynchronous file save 19:51 IhrFussel nerzhul, well Krock tried ~ 10 times to join but always disconnected now while I can normally play 19:52 nerzhul asio offers asynchronous interfaces over a single thread using epoll/kqueue implementation or select on windows 19:52 sofar asynchronous mod code sending to clients :) 19:52 nerzhul sofar, we can imagine asynchronous mapblock save 19:52 sfan5 i'm not sure why we would need that 19:53 nerzhul sfan5, comprssion on another thread ? 19:53 nerzhul permitting to reduce server loop on mapblock save 19:53 nerzhul it's the only case i see 19:53 sfan5 i don't think mapblock serialization is on the server thread rn 19:53 nerzhul it is :p 19:53 nerzhul everything is on serverthread except network and emerge blocks 19:53 sfan5 that makes no sense, who designed that 🤔 19:54 nerzhul packet sending & serialization is on the server thread 19:54 nerzhul i don't know, but it's difficult to change that now 19:54 nerzhul i want to offload some packet verifications parts to network thread too, but i will see that when network will works perfectly 19:54 sfan5 packet sending makes sense kind of, but block serialization really doesn't need to be on the server thread 19:55 sofar ^^ 19:56 nerzhul sfan5, in fact map is only on the server thread 19:56 nerzhul we have a giant lock on it 19:56 nerzhul it's the main difficulty to fix 19:56 sofar max_lag is kinda scary high on my poor VM host... 19:56 nerzhul env_lock is our giant lock 19:56 Fixer *,...,* 19:56 nerzhul if asio is really merged, we can easily port mapblock and other heavy serialization to it 19:57 nerzhul also, on client we have bad design too 19:57 nerzhul poor FPS is generaly due to the fact client & renderer are running on the same thread 19:57 nerzhul if client is slow, FPS are dropped 19:57 sofar %% 19:57 sofar ack ^^ 19:57 nerzhul it's difficult to change that too 19:59 IhrFussel max_lag is not very useful...it goes down too slow so players sometimes think "ahhh high lag" even though there is no high lag right now at all..I suggest to adjust the decrease depending on the recent lag 19:59 sofar max_lag is a measure of how well the server is able to cope with player/mod changes 19:59 sofar so, it's fairly critical that it doesn't spike, or else all players will lag 20:00 IhrFussel Ok now what if my moderators moves something with WE and causes a 10+ secs lag? It will take LOTS of minutes for it to go down again 20:00 sofar it's just an average over 10 minutes 20:00 Fixer IhrFussel: thats okay 20:00 sofar makes it easier to log and monitor 20:00 Fixer IhrFussel: WE can lag for minutes if not crash 20:00 sofar the server admin needs to look at it, not the client 20:01 sofar if it goes up then the clients/players should just pack up and go to a server with a competent admin ;) 20:01 IhrFussel Nah my avg_lag mod is more useful IMO...it logs the max dtime per minute 20:02 sofar url? 20:02 sofar it's just the same data, different algo 20:02 sofar do you globalstep dtime? 20:03 IhrFussel The dtime in Lua is the same as in c++ I think 20:03 sofar do you have your mod code somewhere? 20:04 IhrFussel I can link it but it's not the most efficient code I'm sure 20:04 nerzhul i restart my server to add some logs which can permit to catch Krock problem i think 20:04 sofar doesn't matter 20:05 sofar I ended up printing essentially /status to the minetest log to monitor my server 20:05 sofar so I just do it once every 5 minutes, and since max_lag averages over 10+ minutes, I catch all the high values 20:06 sofar doing a globalstep monitor works too, it's probably fine, it's just more lua ticks per serverstep, but I doubt it hurts much anyway 20:07 IhrFussel sofar, http://ihrfussels-server.tk/lagmod.tar.gz 20:07 nerzhul erf i forgot keepalive option :) 20:08 sofar wow, that's some spartan code, you must hate the space/tab bars 20:09 IhrFussel I work directly on my dedi when I code and it's difficult in nano to work with spaces/tabs 20:09 sofar ow, no test setup? 20:09 sofar I have a full clone of my server where I can test before I deploy 20:10 sofar of course, the map is tiny for insidethebox 20:10 Krock sofar, spaces and tabs are overrated. they just cause longer loading times for Lua, like almost a microsecond 20:11 sofar Krock: data to support that claim, please 20:11 IhrFussel sofar, oh I forgot: You need to remove the avg_lag.critical var as it's not used standalone 20:12 Krock sofar, insufficient data. needs average calculation across most possible mod writers to ensure the result is representative 20:12 IhrFussel I use it to control certain stuff of other mods to "pause" certain features if lag is too high 20:12 Krock the comment wasn't meant serious anyway :P 20:15 IhrFussel nerzhul, Do I need the git am ..... .patch command again? 20:16 nerzhul yeah, type git reset --hard origin/master 20:16 nerzhul and then re-dl the patch and apply 20:17 nerzhul i added tcp keepalives, Krock that should help, i forgot keepalives it's the most important part to ensure socket is always conencts and doesn't timeout :p 20:17 IhrFussel "Server thread should be quiet when no network packet is coming, is means users are inactive" is the correct HEAD? 20:18 Krock ah okay 20:18 nerzhul no it's previous 20:18 nerzhul did you removed the patch and re-dl ? :) 20:18 IhrFussel "4e19791 [CSM] Add callback on open inventory (#5793)" 20:19 nerzhul yeah, you are now on master 20:19 nerzhul then download & apply the patch 20:19 nerzhul or do the Krock suggestion about remote, it should be simple after init :p 20:20 Krock $ git pull nerzhul network_asio 20:20 Krock done .. recompiling 20:21 Krock nerzhul, some warnings for you: https://pastebin.com/raw/nhQckGs0 20:22 nerzhul which compiler Krock ? 20:22 nerzhul i cannot fix warnings from the library :s 20:22 nerzhul it's strange i don't have them on archlinux and FreeBSD 20:22 nerzhul for me the embedded lib doesn't have those, but freebsd package 1.10.6 has others 20:23 Krock gcc (Ubuntu 5.4.0-6ubuntu1~16.04.4) 5.4.0 20160609 20:25 IhrFussel nerzhul, so all those asio:: and void network::ClientConnection::readUDPBody(std::size_t) unused variables are harmless? 20:25 nerzhul i'm on clang, i should test with gcc but i cannot fix that :( 20:25 nerzhul you can ignore, it's because i have a missing check to code client side 20:25 nerzhul session_id will be used :) 20:26 IhrFussel git is really complicated...first try didn't merge the PR it said CONFLICT and created a rebase directory..I had to manually delete that directory, reset to master again and re-apply 20:26 nerzhul git rebase --abort 20:26 nerzhul i think you don't rollback to master 20:27 nerzhul then use Krock's method, it's simple for you after inited 20:27 nerzhul git remote add nerzhul https://github.com/nerzhul/minetest.git 20:27 nerzhul git fetch nerzhul 20:27 nerzhul git checkout nerzhul/network_asio 20:28 IhrFussel nerzhul, it worked on 2nd try is 76600b3 correct now? 20:31 IhrFussel Or is there a git command that tells me what commit I'm on? 20:31 sfan5 git status 20:31 Calinou looks like Medium and Dribbble were merged together, and is now a file hosting service: https://dropbox.design/ 20:32 Fixer lol 20:32 IhrFussel sfan5, that only says "XX commits before origin/master" 20:32 sfan5 oh right 20:33 sfan5 git rev-parse HEAD or git show 20:33 IhrFussel git show is good, thanks =) 20:34 Krock there's also git log, if you want more history and less diff 20:34 IhrFussel nerzhul, seems like it still disconnected me 20:35 IhrFussel Yep join bug still exists too I only see a few mapblocks loaded before everything stops 20:36 IhrFussel Seems to be random, 3rd try loaded normally 20:45 nerzhul or git log :p 20:46 nerzhul IhrFussel, noted, i can't reproduce that, it's difficult 20:46 nerzhul my added errorstream doesn't show what i thought 20:47 IhrFussel I'll spam join now to see how often the disconnect on join bug happens 20:50 IhrFussel 10 tries, bug happened on 3 20:51 nerzhul okay and it's stucked at which point ? 20:52 IhrFussel The map tries to load for a fraction of a second but then everything stops and the player gets stuck with no timeout message 20:52 nerzhul yeah i think there is a transmission problem somewhere 20:54 IhrFussel Happened after 5 minutes now so very random 20:55 nerzhul i shutdowned the server :p 20:55 nerzhul reconnect 20:55 nerzhul i added logs on the waited packet sier 20:55 nerzhul size* 20:56 IhrFussel Want me to spam join again? 20:56 nerzhul you can try 20:56 nerzhul tell me when bug happens 20:57 IhrFussel Huge packet awaited! 20:57 IhrFussel 390155640 20:57 nerzhul nice 20:57 nerzhul it's what i wanted 20:57 nerzhul you have the problem client side, not server side 20:58 IhrFussel Easy to fix? 20:58 nerzhul not easy but it confirm my idea 20:58 nerzhul client is stuck, not server 20:58 nerzhul it will now try to read a packet with 390155640 size 20:59 IhrFussel I didn't get anything now just no loading 20:59 nerzhul in fact you try to load all data in a single packet :) 20:59 nerzhul it's the bug 21:00 nerzhul i restart my server now to add one more log to ensure you are not forced disconnect 21:00 nerzhul you can retry it and tell me 21:01 nerzhul i just want to see if there is a server errro which makes you disconnect 21:01 IhrFussel Try to provoke the bug again now? 21:01 nerzhul please 21:02 IhrFussel Happened 21:03 nerzhul you have the huge packet awaited ? 21:03 IhrFussel Nope nothing 21:03 IhrFussel I'll try again 21:03 nerzhul no no 21:03 nerzhul wait :p 21:04 nerzhul let me a minute to think about this 21:04 IhrFussel Now I see gray again...shut down? 21:06 nerzhul you can 21:06 nerzhul i think i understand the problem i will test on my pc to ensure 21:06 jas_ IhrFussel: just rm -rf minetest and git clone -b network_asio git://github.com/nerzhul/minetest.git each time haha 21:06 jas_ just kidding, but if you clone like that, you can simply `git pull` 21:06 nerzhul jas_, crazy men 21:06 jas_ i do it in a separate directory, minetest-asio 21:06 jas_ git clone -b network_asio git://github.com/nerzhul/minetest.git minetest-asio 21:06 nerzhul IhrFussel, i suspect a concurrency issue and a misssing lock 21:07 nerzhul yeah i forgot a lock on the send queue 21:08 IhrFussel Alright let me know when the pull is ready 21:09 nerzhul yeah both client and server will need to be updated when i finished the fix and test 21:11 nerzhul you can now update 21:11 jas_ that was quick 21:11 nerzhul adding a mutex is not long 21:11 nerzhul and with a very good C++ IDE like clion :p 21:11 jas_ and capable hands! : ) 21:12 nerzhul IhrFussel, wait i forgot to protect a lambda 21:12 IhrFussel Does cmake delete all compiled files when I cancel the process? 21:13 jas_ i don't think so 21:13 nerzhul no 21:13 nerzhul only inprogress 21:14 nerzhul okay it's pushed and i factorized TCP & UDP lambda callbacks 21:14 nerzhul (as they are the same :p) 21:16 IhrFussel PC crashed 21:16 nerzhul heh 21:16 Fixer nice 21:16 nerzhul i think now we are almost perfect, except the socket connection detection (when connecting and when disconnected from server) 21:16 IhrFussel Can I pull and compile now? 21:17 rubenwardy nerzhul, CLion kinda sucks 21:17 rubenwardy it doesn't like operator overloading 21:17 rubenwardy and refactoring fails a lot 21:17 nerzhul operator overloading works perfect 21:18 nerzhul i don't see your issue 21:18 nerzhul and refactoring always works for me 21:18 IhrFussel Can I have the link for the patch again? My bash history didn't save (cause of crash) x.x 21:18 nerzhul i rename classes, and big scope variables in MT it works :p 21:18 nerzhul https://github.com/minetest/minetest/pull/6485.patch 21:18 rubenwardy I always get a type error on overloaded + 21:18 IhrFussel And the merge command was git am -3 patchname ? 21:19 nerzhul rubenwardy, what version ? i have up to date version and never see that 21:19 nerzhul my only issue is std::ios::binary which is unknown 21:19 nerzhul IhrFussel, yes 21:19 rubenwardy 2017.2 21:20 nerzhul 2017.3 was released but i don't have a such problem strange 21:20 nerzhul oh it's overload on + 21:20 nerzhul i never overload +, do we have that in MT ? 21:21 nerzhul yeah 21:21 nerzhul https://lut.im/BD9Ap9lhO7/gWOqzByQSBnOHmP1.png 21:21 nerzhul no problem 21:23 rubenwardy nerzhul, https://i.rubenwardy.com/zlq0NBYpq0.png 21:23 rubenwardy this is on another project 21:23 rubenwardy typedef Vector3 V3s; Vector3 operator+(const T &other) const {} 21:25 nerzhul i think it's due to the typedef and the template 21:25 nerzhul can you report it to clion team ? it should be fixed 21:25 nerzhul i hope they will integrate valgrind at a point 21:25 nerzhul gdb/lldb integration is nice 21:26 rubenwardy I personally compile and run it outside of CLion XD 21:26 rubenwardy as I need to run server and client separately 21:27 nerzhul :) 21:27 IhrFussel nerzhul, cmake jumped from 59 to 100% ... is that normal? 21:27 nerzhul the interesting thing is the breakpoint integration to read memory easily directly in clion, for the entwork debug it was essential to read buffers and see asio and threads easily 21:27 nerzhul IhrFussel, it is 21:29 * Krock wonders what rubenwardy is programming 21:29 IhrFussel First try gray screen >.< 21:29 nerzhul what is your client current commit ? 21:29 rubenwardy Krock, https://blog.rubenwardy.com/2017/08/13/multiplayer-topdown-sandbox-game/ 21:29 nerzhul with pure SFML 21:29 Krock ah, that one :) 21:32 rubenwardy it's going well, just added Lua modding: https://gist.github.com/rubenwardy/a26f579a9e75032034b9a09ae04fc586 21:32 rubenwardy might rewrite the backend to be distributed though, still thinking of how to do that 21:35 Krock how about modifying the playground with Lua modding? 21:35 Krock moving walls, items or similar 21:36 Krock just an idea :) 21:36 rubenwardy yeah, true. There will be a setter at some point 21:41 IhrFussel nerzhul, sorry but happened again ;/ 21:42 nerzhul with the big packet size problem ? 21:42 IhrFussel Nope it only shows the server status and then everything stops loading 21:43 nerzhul what is server status ? 21:43 IhrFussel Sometimes the screen stays just gray and other times part of the map loads 21:44 IhrFussel uptime=1536.39, max_lag=0.239226m clients={nerzhul} 21:45 IhrFussel So it seems like the server doesn't even add the joining player in that case 21:46 nerzhul you are added 21:46 nerzhul currently you are in 21:46 IhrFussel Now I am but before it only told me nerzhul 21:46 nerzhul i see join/leave each time in the logs :p 21:46 IhrFussel Now it happened again 21:46 nerzhul the problem is we have a reading issue somewhere or a data race somewhere in the code 21:47 nerzhul 2017-10-04 23:46:38: ACTION[Server]: IhrFussel joins game. List of players: nerzhul IhrFussel 21:47 nerzhul 2017-10-04 23:46:38: ACTION[Server]: IhrFussel leaves game. List of players: nerzhul 21:48 IhrFussel Whatever it is is must be linked to joining and map loading 21:49 IhrFussel Now it's gray again but I see your names 21:50 nerzhul it's because you have many packets at the same time at this moment 21:50 nerzhul and the concurrency issue is trigger, but only for you 21:51 Krock nerzhul, memory leak just occured again. any server logs? 21:51 IhrFussel Could that also have caused the bad_alloc() error a few hours ago? 21:51 nerzhul Krock, memory leak ? 21:51 Krock if not, then it's entirely client-sided 21:52 Krock nerzhul, sudden memory jump from 140 MiB to some GiB 21:52 nerzhul Krock, ouch, it can be interesting to see where is the memory with valgrind 21:52 Krock well.. the system hangs somewhat in that case :< 21:53 Krock it happens in-game and isn't related to connecting or disconnecting 21:54 Krock I suspect a wrong packet size but could be entirely wrong there 21:55 nerzhul i think it's the problem Krock 21:56 nerzhul i just don't see where is the concurrency problem atm 21:56 nerzhul it should occurs on this specific problem 21:58 IhrFussel RAM usage is constant on my side currently 21:58 nerzhul yeah me too 21:59 IhrFussel NVM I got the error 21:59 nerzhul i will look at this part tomorrow, i'm tired, thanks for your tests :) 21:59 IhrFussel From just under 1 GB to 1.5 GB in a sec 22:00 IhrFussel Huge packet awaited, size: 485636303 22:02 nerzhul 485636303 = 460MB of allocated memory 22:02 nerzhul seems to be linked :) 22:02 nerzhul see you tomorrow, thanks 22:04 Fixer very nice work 22:05 Krock IhrFussel, that's like you'd order a ship in a regular amazon package 22:07 Krock expected material much > than what you've got 22:09 Krock somehow similar to http://i.imgur.com/x4TnRDC.png 22:10 jas_ rubenwardy: that game (and video) looks sooo cool 22:10 rubenwardy thanks :D 22:10 jas_ good luck with it! 22:14 Fixer nice troll 22:20 shivajiva http://www.commitstrip.com/en/2013/03/07/des-bonnes-conventions-de-nommage/ 22:27 Fixer qtpfsgui 22:34 sofar whenever I use that thing, I can never figure out how to make the photographs look like a professional did them... too many options, most if not all are garbage algorithms 22:52 Fixer "git clone -b network_asio git://github.com/nerzhul/minetest.git" 22:52 Fixer saved 22:53 Fixer provocative offtopic question 22:53 Fixer is your religion allows you to use ! in filenames? 22:56 jas_ git clone -b network_asio git://github.com/nerzhul/minetest.git minetest-asio <--- saves to minetest-asio directory 22:57 jas_ i;m allowed to use ! in filenames, yes 22:57 Fixer i kinda feel uneasy 22:57 jas_ touch \! 22:57 Fixer but it is useful 22:58 jas_ rm \! 22:58 Fixer i feel good with _name.ext but !name.ext feels uneasy 22:58 Fixer maybe it will not work in other FS or on CD/DVD/whatever 22:59 Fixer compiling 23:14 Fixer still having that "client disconnected" problem 23:44 ircSparky has anyone noticed that copy and pasteing with 0.4.16 books dosnt work with the page system? 23:44 ircSparky (probably, just wondering)