Time Nick Message 00:23 laudecay have any of you heard about efforts to develop a p2p sort of framework on top of this? my research group is thinking about forking your code and rewriting the engine to add deterministic lockstep and other cool P2P algorithms 00:33 sofar laudecay: p2p minetest? what parts would you peer? the map data? 00:33 rubenwardy lockstep isn't well suited for this sort of game 00:33 sofar laudecay: in general MT is server centric due to the requirements that players can't just make unauthorized changes 00:34 rubenwardy lockstep is good for RTSes 00:41 red-001 #6951 00:41 ShadowBot https://github.com/minetest/minetest/issues/6951 -- [CSM] Don't create the script environment if CSM is disabled by red-001 05:03 paramat celeron55 please could you consider my suggestions in https://github.com/minetest/minetest/issues/5393#issuecomment-359675913 and comment? 05:30 paramat merging #6932 05:30 ShadowBot https://github.com/minetest/minetest/issues/6932 -- Remove CSM screenshot API by red-001 05:38 paramat merged 06:46 paramat merging game#2015 game#2020 game#2024 06:46 ShadowBot https://github.com/minetest/minetest_game/issues/2015 -- default:dirt_with_snow: Re-add to soil group by paramat 06:46 ShadowBot https://github.com/minetest/minetest_game/issues/2020 -- Creative: New textures for less distortion by paramat 06:46 ShadowBot https://github.com/minetest/minetest_game/issues/2024 -- Dungeon loot: Reduce maximum processed rooms to 8 by paramat 06:51 paramat merged 14:42 ThomasMonroe #6955 14:42 ShadowBot https://github.com/minetest/minetest/issues/6955 -- Update tab headers by ThomasMonroe314 14:51 ThomasMonroe red-001, are you going to work on #6048? 14:51 ShadowBot https://github.com/minetest/minetest/issues/6048 -- Add a refresh button to the server list. by red-001 14:59 red-001 ThomasMonroe, not really, not great at user interface design, so if you want to make an icon for it feel free 14:59 ThomasMonroe icons don't actually work in the mainmenu for some reason 15:00 red-001 huh 15:00 red-001 you sure? 15:00 ThomasMonroe the images don't show up at all 15:00 red-001 so how do the subgame icons work? 15:01 ThomasMonroe idk, those aren't lua based as far as I can tell 15:01 red-001 pretty sure they are 15:01 red-001 had to move them for something 15:01 ThomasMonroe at least defining them that is 15:02 ThomasMonroe I can't see anything in the tab_local.lua that would put them there 15:04 red-001 we really need to get rid of that android menu one way or another 15:04 red-001 either it's too large to work anyways or it locks android users out of features 15:05 red-001 Jordach, you were working on redesigning the menu right? 15:05 ThomasMonroe I'll pick up #6048 if you don't mind me doing that 15:05 ShadowBot ThomasMonroe: Error: Delimiter not found in "Connection timed out." 15:06 ThomasMonroe oops 15:06 ThomasMonroe wait what? 15:06 ThomasMonroe oh nvmd 15:12 red-001 if you want to, sure 15:13 ThomasMonroe ok 15:37 pgimeno_ I like the idea of 6048. Preferably, with an option to disable refreshing the list by default. That way, when I am going to play locally, I don't have to wait for the list to fill up, and the server also receives fewer requests. 15:38 ThomasMonroe http://imgur.com/Dvz98N1l.png how's this pgimeno_, red-001 15:39 rubenwardy > serveminetest.net 15:42 pgimeno_ lgtm :) (but take my opinion as that of just a user) 15:44 ThomasMonroe ok, here's the PR: #6957 15:44 ShadowBot https://github.com/minetest/minetest/issues/6957 -- Add a refresh button to the serverlist by ThomasMonroe314 15:49 pgimeno_ I guess the same objection as in 6048 will apply 15:51 ThomasMonroe about the image button? 15:51 ThomasMonroe they don't work in the mainmenu for somereason 15:51 red-001 -1 unless it gets a image button 15:52 red-001 that was the whole reason the first PR was put on hold 15:54 pgimeno_ 'strace -e trace=open minetest' may help finding out what is trying to open and why it doesn't work 15:54 pgimeno_ (the image in the menu) 16:00 ThomasMonroe i would put a Unicode char in there, but idk how to get Lua to do that 16:02 pgimeno_ if it's supported, it will most likely be UTF-8 16:04 pgimeno_ ↻ maybe? 16:05 ThomasMonroe yes!! 16:05 ThomasMonroe what char was that? 16:06 pgimeno_ that's U+21BB, encoded as \xE2\x86\xBB in Lua 16:06 pgimeno_ in UTF-8 16:08 red-001 #6954 16:08 ShadowBot red-001: Error: Delimiter not found in "Connection timed out." 16:20 ThomasMonroe erm idk if this version of lua supports unicode 16:24 ThomasMonroe er anything higher than UTF-8 that is 16:27 red-001 #6954 16:27 ShadowBot https://github.com/minetest/minetest/issues/6954 -- Remove unused packet IDs and network code. by red-001 16:27 red-001 Megaf, it just breaks sometimes 16:28 Megaf well done red-001, on the cleaning up 16:28 red-001 ~1% of the time it just fails to get the url 16:28 Megaf LGTM 16:39 ThomasMonroe ugh pgimeno_, it doesn't display the char in the main menu 17:26 ThomasMonroe !tell Krock I can't get the unicode char ↻ to show up in Lua 17:26 ShadowBot ThomasMonroe: O.K. 17:27 red-001 your daily reminder that #6940 still needs a review 17:27 ShadowBot https://github.com/minetest/minetest/issues/6940 -- Fix issues with earlier CSM HUD commit by red-001 17:32 ThomasMonroe hiya Krock 17:59 lisac nerzhul: can you please review #6946? I've got one approval already. 17:59 ShadowBot https://github.com/minetest/minetest/issues/6946 -- Disable fall damage when "immortal" group set by lisacvuk 18:00 nerzhul lisac, done 18:04 nerzhul merging #6952 18:04 ShadowBot https://github.com/minetest/minetest/issues/6952 -- Intersects_protection(): Move from Minetest Game to builtin by paramat 18:04 lisac yay thanks nerzhul 18:05 Krock nerzhul, while you're at it.. mind doing a 2nd pass on #6787 ? 18:05 ShadowBot https://github.com/minetest/minetest/issues/6787 -- textarea[], field[]: Unify function, fix wrong fallback text by SmallJoker 18:11 nerzhul Krock, erf, i should hide :D 18:12 nerzhul Krock, code is good but i didn't tested it 18:12 * Krock watches nerzhul quickly hiding behind a tree 18:14 Megaf software video renderer totally broken, should I open an issue about it? https://imgur.com/a/LaO4d 18:14 Krock yes, but not on minetest/minetest 18:15 Krock time for irrlicht devs to fix their stuff 18:15 nerzhul Krock, are you still okay for #6946 ? 18:15 ShadowBot https://github.com/minetest/minetest/issues/6946 -- Disable fall damage when "immortal" group set by lisacvuk 18:16 Megaf Krock: it's time for Minetest to stop using this crap 18:16 Megaf yet another bug that we blame on Irrlicht 18:16 sfan5 "we blame on Irrlicht" what? 18:16 sfan5 it is literally Irrlicht's fault 18:17 Megaf +1 reason to not use Irrlicht 18:17 sfan5 not really, the software renderer is not intended for serious use 18:17 Megaf we can just use TuxKarts engine 18:17 sfan5 is this a joke 18:17 red-001 hey I found a way to fix it 18:17 red-001 don't use the software render 18:18 nerzhul red-001, CSM particles are nice, it permits to have local particle which doesnt' polute the server loop everywhere on the map, localized on players :p 18:18 Krock nerzhul, trivial changes all over the place. sure it's still fine 18:18 nerzhul Megaf, i tried to change MT to STK irrlicht and it's not very possible has they heavily change it 18:18 nerzhul red-001, seems your fix is like removing CSM to fix it :D 18:19 red-001 huh which one? 18:19 nerzhul Krock, then you have my approval go ahead 18:19 nerzhul red-001, i do a similar comment as software renderer fix you proposed :D 18:20 Megaf nerzhul: it's a shame, it does look quite good https://supertuxkart.net/Antarctica:_Technical_Details 18:20 Megaf lots of nice features 18:20 Megaf and working occlusion 18:20 nerzhul maybe but the API is broken :p 18:20 nerzhul and it's a lot of work 18:21 nerzhul i strrted to prepare a such change with many client refactoring, but without properly spliting renderer from client it's very hard to change it 18:21 red-001 yeah architecture is a mess 18:21 nerzhul we have many irrlicht call everywhere in client in many objects 18:22 nerzhul abstracting some changes is needed to make it happen 18:22 Krock merging #6946 and trivial #6953 in ~5 minutes 18:22 ShadowBot https://github.com/minetest/minetest/issues/6946 -- Disable fall damage when "immortal" group set by lisacvuk 18:22 ShadowBot https://github.com/minetest/minetest/issues/6953 -- Add languages missing from settings menu by MuhdNurHidayat 18:22 Megaf nerzhul: what if I just try to build Minetest pointing the irrlicht lib to this directory? https://github.com/supertuxkart/stk-code/tree/master/lib/irrlicht 18:23 Megaf and maybe updating the defines stuff 18:23 Megaf wouldnt it work? Or thats what you tried 18:23 sfan5 Megaf: that source directory is not the "Antarctica" engine the wiki page mentions 18:24 red-001 nerzhul, my point with removing obsolute op-code is to make it easier for people of the future trying to find a new free id 18:25 red-001 marking them as obsolete makes it seem like they shouldn't be reused 18:25 nerzhul red-001, we are using less than 10% than the possible ids no ? opcodes are on 16b 18:26 rubenwardy nerzhul, Antarctica is a fork of Irrlicht 18:26 nerzhul rubenwardy, it doesn't mean they keep compat, last time i tried to move to it (2 years ago) MT was totally broken, but API was identical 18:26 rubenwardy so at worse it'll be like changing major versions of Irrlicht - some API breakages 18:27 nerzhul maybe it's better now, but no distro embed it, then we should have a submodule or a copy in our repository 18:27 Krock merging the PRs now... 18:28 lisac two approvals is enough for a merge, right? :) 18:28 lisac hey Krock 18:28 lisac yay 18:28 * lisac has another bugfix merged! 18:29 red-001 nerzhul, huh 18:29 TheBugHunter lisac: o/ 18:29 TheBugHunter congrats 18:29 red-001 I thought there was a lot less 18:29 lisac thanks :) 18:30 red-001 so TOSERVER_NUM_MSG_TYPES isn't the max number of opcodes? 18:30 rubenwardy yes 18:31 rubenwardy it's a idiom to automatically get the number of op codes 18:31 Krock depends on what integer type is used to detect that 18:31 TheBugHunter [18:27:03] maybe it's better now, but no distro embed it, then we should have a submodule or a copy in our repository 18:31 rubenwardy so it must be at the end of the enum list 18:31 TheBugHunter I see now issues there 18:31 TheBugHunter we bundle already jsoncpp, jmp and sqlite3 18:31 TheBugHunter gmp* 18:32 Krock irrlicht is quite heavy and requires other libraries too, do you know that? 18:32 Megaf I think we should NOT bundle jsoncpp and GMP, and bundle instead a/our fork of Irrlicht/other engine 18:32 Megaf And we don't need the whole thing... 18:32 red-001 I suppose I just kinda assumed with all the free space in the packet table the size was hardcoded 18:33 nerzhul Megaf, it's ridiculous to embed irrlicht and not the common libs between server and client 18:34 rubenwardy it's u16 max I believe 18:34 nerzhul embedding jsoncpp and sqlite3 permit to have those librairies up-to-date and stable, and they are tiny 18:34 nerzhul rubenwardy, theorically yes 18:34 red-001 rubenwardy, wow 18:35 rubenwardy tbh, once we reached 1000 0000 we could have made it expanding 18:35 rubenwardy ie: a first bit of 1 means the next byte is used 18:35 rubenwardy nvm, that's unnecessary 18:38 red-001 done #6954 18:38 ShadowBot https://github.com/minetest/minetest/issues/6954 -- Remove legacy chat network code. by red-001 18:39 Krock as if we ever will need that many different packets 18:39 rubenwardy 64MB is another for anyone! 18:39 rubenwardy in all serious, getting more than 255 would be suprising 18:41 Krock relevant xkcd: https://www.xkcd.com/619/ 18:41 red-001 !title 18:41 ShadowBot xkcd: Supported Features 18:42 Megaf dont know if I should work on forking Irrlicht and making a tiny version of it for Minetest or porting Minetest to Antartica 18:42 sfan5 you don't need to fork Irrlicht if you want a tiny version 18:42 sfan5 there's lot of optional features you can disable by editing irrCompileConfig.h 18:43 Megaf sfan5: I mean, to bundle it in Minetest 18:43 sfan5 we won't bundle Irrlicht, so that would be wasted effort 18:45 Megaf you could bundle µLight, that'd be just hte necessary libs we need 18:46 Megaf (Micro)µLight = tiny version of Irrlicht 18:46 sfan5 but we won't 18:46 Megaf why not? What's so bad in bundling an essential lib and having control over it? 18:47 rubenwardy more maintenance 18:47 rubenwardy minimal gain 18:48 Megaf What is GMP used for by the way 18:48 rubenwardy Minetest is 20MB all in all 18:48 Megaf it's a Linux mouse lib for the terminal, text mode 18:49 Megaf ah 18:49 Megaf lol 18:49 rubenwardy that's including DLLs, quite often you easily reach that without them 18:49 Megaf its a different gmp 18:55 rubenwardy web#111 18:55 ShadowBot https://github.com/minetest/minetest.github.io/issues/111 -- Update customize page by rubenwardy 19:00 rubenwardy Calinou: any objections to web#79 ? 19:00 ShadowBot https://github.com/minetest/minetest.github.io/issues/79 -- add new background-downloads.jpg by SpaghettiDeveloper 19:04 Calinou rubenwardy: looks good to me, I can merge it if you want 19:04 rubenwardy I can do it now, have a trivial one too 19:06 rubenwardy Calinou: https://github.com/rubenwardy/minetest.github.io/commit/f42be28dced5cc5153413dc431b18179237b0ee1 19:06 rubenwardy also, we should change from GNU/Linux to Linux 19:07 rubenwardy only stallman's fanatics care about the ugly name 19:07 rubenwardy 19:07 red-001 idk I wouldn't risk messing with fanatics 19:08 red-001 plus I hear he was a katana and it's a well known fact that they are the pinnacle of swordsmanship 19:08 red-001 s/was/has but both work 19:09 sfan5 rubenwardy: agree 19:10 Krock I'd like to interject for a moment 19:11 nerzhul i just pushed a SSM api change #6958 19:11 ShadowBot https://github.com/minetest/minetest/issues/6958 -- Add minetest.bulk_set_node call + optimize Environment::set_node call by nerzhul 19:11 nerzhul (yeah i should work on CSM but i got this idea before :p) 19:11 nerzhul and i got 10 mins to cod ethat 19:12 Krock wrong indents in the docs 19:12 nerzhul too fast for me Krock 19:12 nerzhul but yes :p 19:12 Krock missing + z in line 699 19:12 red-001 huh 19:12 nerzhul no it's normal :p 19:12 red-001 I wonder could we rewrite tnt to use this 19:12 nerzhul i didn't wanted to do a cube :p 19:13 red-001 nvm looks like it wouldn't work 19:13 nerzhul tnt use remove node no ? maybe if this call is pleasant for modders i will push similar thing for remove node :p 19:13 nerzhul but if your prefer a cube in integration test i'm neutral xD 19:13 nerzhul Krock, doc fix pushed 19:13 Krock nerzhul, then your loops make no sense as you don't need the z iteration 19:14 red-001 nerzhul, tnt needs to also read some data, so it needs full voxelmanipl 19:15 nerzhul Krock, exact :p 19:15 rubenwardy merging... 19:15 rubenwardy done 19:16 nerzhul Krock, you achieve me, okay for a cube :p 19:17 Krock :P 19:17 nerzhul and then, i triggered 4 travis builds yeah :p 19:17 red-001 anyone have anything agaist moving chat command handling to core? It would allow autocompleting commands and configuring the command prefix 19:17 nerzhul macosx users will be happy haha 19:17 nerzhul the main registration table ? 19:17 Krock you can abort them easily. usually it has a delay of a few minutes to start running 19:17 red-001 Krock, s/minutes/hours 19:18 red-001 at least for max os 19:18 Krock s/max /mac/ 19:18 red-001 s/ // 19:19 red-001 oh nvm they changed the name again 19:19 Krock nerzhul, sorry that I haven't seen it in the first pass but there's a performance improvement possible. reserve the pos_list vector length to `len` 19:20 nerzhul Krock, yeah it's exact i missed that part, i was focused on the lua improvement 19:20 Krock `bool s` is also redundant 19:20 nerzhul exact 19:21 Calinou rubenwardy: that's ok 19:22 red-001 nerzhul, move detecting commands being issued and the help command to the client, and add tab autocompletion for commands 19:22 nerzhul Krock, done heh 19:22 red-001 on the server side add a new on_chatcommand callback 19:22 nerzhul red-001, then you will send available commands to client ? 19:23 red-001 yes 19:24 red-001 add a new internal api that registers a command and all the data needed for /help and make the chat command code call that 19:24 red-001 kinda like register node 19:29 nerzhul red-001, if it's exactly the current API should be good 19:38 rubenwardy Calinou, web#112 19:38 ShadowBot https://github.com/minetest/minetest.github.io/issues/112 -- Move home sub items to root level by rubenwardy 19:57 Megaf uh, help? Linking failing in 0.5 from master. http://paste.debian.net/plain/1006678 19:57 sfan5 -DENABLE_SYSTEM_JSONCPP=1 19:57 sfan5 this keeps happening but we have no idea why so we can't fix it 19:58 Megaf eh, what about was said before about bundling json? 19:58 Megaf :) 19:58 Megaf bundling 19:59 Megaf sfan5: it only happens on 0.5, and I think it only happens on Debian Jessie, it works on Debian Stretch 19:59 Megaf Stretch* 19:59 sfan5 dunno what was said? 19:59 Megaf [18:34:23] embedding jsoncpp and sqlite3 permit to have those librairies up-to-date and stable, and they are tiny 19:59 Megaf :D 19:59 Megaf I say we drop json 19:59 sfan5 well, it's up-to-date and stable 19:59 sfan5 just broken on debian 20:09 Wayward_One fwiw, i experienced that on linux mint too 20:15 Megaf sfan5: even worse with systemjson http://paste.debian.net/plain/1006682/ 20:16 sfan5 your system json is too old 20:16 Megaf It's Debian 8, just one of the most popular Linuxes... 20:16 Megaf MT is too "advanced" 20:16 sfan5 wrong 20:16 sfan5 debian is too old 20:17 sfan5 nerzhul: can you try to find a version of jsoncpp that compiles & links fine on debian 8? 20:17 pgimeno_ I'm on Debian 9 20:17 sfan5 this issue is getting quite annoying and it would be better to just fix it (even if we have to downgrade jsoncpp) 20:18 nerzhul sfan5, what do you mean, the repository version is not good for 8 ? 20:18 sfan5 yes, it fails to link http://paste.debian.net/plain/1006678 20:18 Megaf sfan5: to be clear, I'm fine with running Debian 9 and supporting Debian 9. I just thing it would be interesting supporting Debian 8 too due to it's widespread use 20:19 sfan5 nerzhul: this keeps happening: https://github.com/minetest/minetest/issues/6567#issue-269361883 20:20 nerzhul downgrading jsoncpp means less performance and some bugs can happen 20:21 sfan5 I'd rather have slower json parsing than broken build on debian 8 20:21 nerzhul use system json cpp then 20:21 nerzhul it's the prefered version for old systems 20:21 sfan5 that seems to be too old to work 20:21 nerzhul wtf 20:22 nerzhul https://packages.debian.org/fr/jessie-backports/libjsoncpp-dev 20:22 nerzhul use backports 20:22 nerzhul don't use main repository ,use backports repository and it's nice 20:23 nerzhul i have debian 8 only servers at work (yeah i'm late too) and we heavy use backports for many softwares (mariadb, apache, haproxy, etc...) it works like it should 20:23 nerzhul 0.6 in main repository omg 20:27 Fixer Megaf: not sure if 8 is supported, upgrading to 9 should not be a problem, unless the are some problems 20:27 Megaf nerzhul: how faster can new json make minetest? 20:28 Megaf Maybe just as fast as removing all mapgens and dbs and leaving just sqlite3 and v6 20:28 rubenwardy so not faster at all 20:29 Megaf that's my point 20:34 Megaf systemjson wasnt linking on debian 9 either, Im not 100% sure if this was fixed #4306 20:34 ShadowBot https://github.com/minetest/minetest/issues/4306 -- Minetest fails to link JSONCPP on Debian 9.x if -DENABLE_SYSTEM_JSONCPP=1 20:35 pgimeno_ ah yeah I have it off 20:39 Megaf enabling backports here... 20:43 nerzhul rubenwardy, 10-12% perf enhancement in debug mode 20:49 nerzhul same in relwithdebinfo 20:49 nerzhul retrying release 21:08 nerzhul 25% less 21:08 nerzhul it's nice 22:04 Megaf nerzhul: ok, updated everything to backports. Lets see if it builds and links using system jsoncpp 22:06 Megaf Thou the begining was not very promissing https://i.imgur.com/LFWfpa6.png 22:07 Megaf wait, hold on 22:07 Megaf -- Using bundled JSONCPP library. 22:07 Megaf lol, 22:11 Megaf Sorry, I thought using systemjson was the default if the lib was found. Compiling again with the use systemjson flag 22:18 Megaf nerzhul: you sure #6958 will not break anything? 22:18 ShadowBot https://github.com/minetest/minetest/issues/6958 -- Add minetest.bulk_set_node call + optimize Environment::set_node call by nerzhul 22:19 Megaf And did you test the memory spike when running the command with a large number of nodes, lets say, 100000 22:20 nerzhul no overhead 22:20 nerzhul i tested on 100 * 100 *100 nodes 22:21 Megaf I want to see how does that compares/relates to worldedit 22:21 Megaf and if worldedit could make a good use of that 22:21 nerzhul 12MB C++ memory for the for loop it's the only overhead for the price of 25% less time processing on 1M nodes 22:21 nerzhul worldedit can without doubt use it, but i think we can add better calls for worldedit that this generic call 22:24 ThomasMonroe nerzhul, why shouldn't I capitalize "Game" ? I thought that would be consistent with the other tabs. 22:24 sfan5 i don't think this is very useful for worldedit 22:24 sfan5 most commands use vmanips anyway 22:38 Megaf About JSON, it linked using systemjson with debian 8 backports 22:38 Megaf also linked using bundled I think 22:39 Megaf I'm building client and server, minetestserver has linked already, client stil building 22:58 pgimeno_ compiled just fine on Debian 9. -- Using system JSONCPP library. 22:59 pgimeno_ that was without the standalone server 23:10 Megaf pgimeno_: problem is with Debian 8 23:20 pgimeno_ Megaf - 4306 was about Debian 9, you said you were not 100% sure if it was fixed 23:22 Megaf Oh, thanks