Time Nick Message 00:08 paramat actually i'm not sure updateLighting is buggy, got confused 00:12 paramat ah but the 'update map' api also calls updateLighting, and 'update map' causes shadow bugs 00:20 est31 okay lets talk about VAEs again, and how to store them inside the map 00:21 est31 does a 32 bit realm number seem apropriate? 00:21 est31 16 bits are a bit few, if we want to use VAEs inside mapgen alot 00:23 paramat realm number = max number of VAEs? 00:23 est31 yes 00:23 est31 imagine making sky islands VAEs 00:25 paramat for storage how about reducing world size to +-30000 and using the 1000 deep border? 00:25 est31 I dont like that 00:26 est31 its an object at the other end of the map 00:26 est31 I'd rather store them manually as schematics 00:26 est31 inside an extra dir 00:26 paramat ah 00:27 est31 But about doing it properly, I'd suggest to add a new column to the map table 00:27 Sketch2 is anybody else getting strange visual effects from the latest minetest on windows? 00:27 paramat i was confused by 'how to store them inside the map' 00:27 est31 perhaps having a map table, and a second VAEmap table 00:28 Sketch2 it seems to be a couple of my friends are getting this strange effect as if things are being highlighted that aren't being pointed to 00:28 Sketch2 https://www.dropbox.com/s/k4fkl6a15cvk19s/Screenshot%202015-08-18%2000.14.51.png?dl=0 00:28 est31 opengl 2.1 00:28 Sketch2 I know of two people that are using laptops with Intel graphics that this shows up on 00:29 est31 are shaders enabled? 00:29 Sketch2 yeah, 2.1. either way with the shaders. 32 or 64 bit same thing. 00:29 * Siva_Machina forgot I was lurking in here 00:29 est31 I guess its the same as #2866 00:29 ShadowBot https://github.com/minetest/minetest/issues/2866 -- Graphical regression. 00:29 Siva_Machina >.> 00:30 est31 can you upload the screenshot there via drag+drop? 00:30 paramat 'node highlighting' enabled? 00:30 Siva_Machina no node hightlighting was not enabled 00:31 Siva_Machina tried turning off the shaders 00:32 Siva_Machina https://photos-1.dropbox.com/t/2/AAA6VEBelLrdCbVddmu41l4QNQvGWSeMjJFHf56M4tFhkQ/12/393874156/png/32x32/1/1440208800/0/2/Screenshot%202015-08-18%2000.14.51.png/COyV6LsBIAEgAiADIAQgBSAGIAcoAg/FlzGanhibycQKFG_jeG5enmykXxIj03XjXsq9Fm65M0?size=1024x768&size_mode=2 00:33 Siva_Machina that screenshot is from a different build of 0.4.11 but I have the same issue on 0.4.13 00:34 est31 then its not 2866 00:34 Siva_Machina it looks like it is 00:35 est31 2866 appeared since 0.4.12 00:36 est31 due to some changes by RBA 00:36 Siva_Machina OK...that makes since. Since I ran into in a build of 0.2.11.....but in 0.4.12 I do not get that issue with or without the shaders 00:37 est31 also your screenshot appears to be from 0.4.12 something 00:37 est31 not 0.4.11 00:38 Siva_Machina ah 00:38 Siva_Machina might have been...xD 00:38 Siva_Machina <.< 00:38 est31 the minetest version of the screenshot is from after RBA's changes. 00:38 Siva_Machina anyways the two main build of Minetest 0.4.12 do no have it 00:39 Siva_Machina have that issue* 00:39 est31 then its 2866 00:40 Siva_Machina most likely 00:53 paramat anyway i agree best VAEs are not stored in the world somewhere 00:53 est31 we'll already need some other place to store other intermediate data 00:54 est31 to make it possible to manage large VAEs 00:54 est31 imagine you have a huge spaceship, and its (0,0,0) is inside an unloaded area 00:55 est31 you'll need to know to load it when you are close to its hull 00:55 est31 fortunately its the same problem as the areastore problem 00:55 est31 so we have a data structure for it 00:55 est31 we just need to do the serialisation 01:04 est31 paramat, is the fractal mapgen pr final? 01:04 paramat no, it's usable but in development 01:05 paramat i just got fed up with manual rebasing, easier to dev it while it's in 01:06 OldCoder Basic question, refer me to docs if appropriate, why are MT colors on my display a bit washed out? Looks a lot better than I up contrast on screenshots 01:09 est31 okay 02:52 hmmmm est31: does 0.4.13 compile on android? 02:53 est31 yes it does 02:53 est31 just it makes a wrong version name 02:53 hmmmm what was kahrl saying about how wctomb not compiling or something because 'tmp' is undefined in a function 02:53 hmmmm ? 02:54 est31 that was because of git weirdnesses 02:54 hmmmm ahh 02:54 hmmmm okay good 02:55 hmmmm so at this point are we ready to continue on with 0.4.14 development or are there pressing matters we need to address right now? 02:55 est31 no, release of 0.4.13 went fine 02:56 hmmmm okay just checking 02:56 hmmmm hehe it feels good to not be in a freeze 02:56 est31 the only thing that makes me sad is that TBC_x's name has been gotten wrong 02:56 hmmmm well 02:56 hmmmm sorry about that 02:56 est31 yeah, everybody makes mistakes 02:56 hmmmm hey did you see there was a guy who made a minetest lua api unittest 02:56 hmmmm mod 02:58 hmmmm i wonder if we can add that to minimal 03:03 est31 you mean https://github.com/t4im/mtt 03:03 est31 yea good idea 03:03 hmmmm yeah that one 03:03 est31 but t4im maintains it 03:03 hmmmm shrug 03:04 est31 if we add it to minimal, he can't do that anymore 03:04 est31 also mtt uses require alot 03:04 hmmmm maybe we can sync our repository's copy every month or so 03:04 est31 which breaks with require 03:04 est31 err 03:04 est31 mod security 03:04 hmmmm ahh 03:06 est31 perhaps we could set the CI bots up to sync with his repo 03:07 hmmmm hmm 03:07 est31 so that every PR gets tested after mtt 03:07 hmmmm unittests should really be maintained by the developers, not a third party 03:07 hmmmm i mean what if you change the output of a lua api wittingly and that makes it fail the unit test 03:08 est31 lets wait what t4im says to the idea 03:09 est31 disclaimer: he's a friend of mine 03:10 hmmmm as far as making the ci bots depend on the unittest results... i don't see how that's happening 03:10 hmmmm many of the api effects require human judgement 03:11 est31 at least we can have regression tests 03:18 est31 there he is :D 03:23 T4im well, not sure how much it could be integrated into ci either, but mostly, because it still requires to run from within minetest… have experimented a bit with getting stuff run without the engine, but less for the api testing, and more for the mod testing part (and it didn't go well enough yet) 03:24 est31 you can run arbitrary shell commands on the CI 03:25 hmmmm i guess we can create a command line option to run lua unittests 03:25 est31 it is enough to pull the project from git, add it as mod to minimal, and then start a world with minimal 03:25 hmmmm it would set up a mock Server and execute an on_unit_test() callback which will be the entry point 03:25 est31 it then creates a file "success" in the dir 03:25 hmmmm or something like that 03:26 T4im currently it doesn't fail hard btw, so I guess you would want the server to throw an error for negative return value then? 03:26 est31 local a = "fake" 03:27 est31 a() 03:27 hmmmm yeah, isn't there an error() function? 03:27 T4im the whole thing is designed to continue with further tests, or I wouldn't be able to do anything aymore.. many of the bugs are open since half a year :D 03:28 T4im well "many", there aren't many test cases yet.. 03:30 hmmmm gonna start looking into them 03:30 hmmmm i feel like there isn't enough of an emphasis on solving bugs 03:31 hmmmm previously people just cared about their shiny new features, that needs to change though 03:31 T4im original plan was more to get a unit testing for mods available, since that was missing, but well.. then you start stumbling over bugs, and well.. why not write tests for that as well 06:41 est31 Impressive bounty amount: https://github.com/minetest/minetest/issues/3085#issuecomment-133644602 06:58 nore est31: this is *a lot* 06:58 nore although, I understand why 07:00 est31 why do we have a bounty label 07:00 est31 I havent added it 07:00 est31 and only labeled that issue because we have a bounty label 07:01 nore est31: I remember it was added some time ago (when Kenney made bounties) 10:20 TBC_x who did the hard work and commented methods of ClientInterface? 10:21 TBC_x i mean those 10:21 TBC_x /* create client */ 10:21 TBC_x void CreateClient(u16 peer_id); 10:21 Donillo haha 10:22 est31 agreed this kind of comments is stupid 10:22 rubenwardy git blame helps 10:23 TBC_x sapier 10:23 Donillo yea 10:23 Donillo :D 10:26 TBC_x looks like ClientInterface was broken since day 1 10:28 nrzkt yes, the ClientInterface design is a little bit redundant with PlayerSAO & RemotePlayer, this is ... hard. 10:31 TBC_x it must be redesigned, I would strive to data driven approach 10:32 TBC_x and remove client state, at least its present form 10:32 TBC_x commit e258675eabc874d31bc9c6cf49e4bbc1f7f3f417 10:32 TBC_x -add client states to avoid server sending data to uninitialized clients 10:33 est31 the state gives a good overview over what happens on the server 10:33 est31 replaying everything with bools is actually an antipattern, if you believe a rust linter project that is built to point precisely this out 10:34 TBC_x I would simply not create an unitialized client 10:35 est31 its not just a bool, look at the state diagram 10:35 est31 the state manages log in stuff, and more 10:35 TBC_x yeah 10:35 TBC_x that shoud be divided 10:35 est31 and such a diagram is much easier to understand than ten bools 10:36 TBC_x I don't intend to use bools 10:36 est31 where you dont know which is true when the other is true as well 10:36 est31 TBC_x, tell me about your intended concept 10:37 TBC_x present client state holds states of two different things 10:37 TBC_x auth_state and init_state 10:38 TBC_x and I don't like the way of how there is used lt, gt comparison against states 10:38 est31 its called hierarchy 10:39 est31 we can abandon it if we find an additional state that doesnt need it 10:40 est31 err 10:40 est31 if we find an additional state that doesnt work with this hierarchy. 10:41 est31 heard of CPU rings? 10:42 est31 same model here, just turned around 10:42 est31 so right now we have one dev who likes it and one dev who doesnt. 10:42 kilbith it's only me or there's a rendering regressionhttps://play.google.com/store/apps/details?id=multicraft.net that freeze the game for few ms to ~0.5s ? 10:42 kilbith oops 10:42 kilbith -URL 10:43 TBC_x I doubt that it is comparable to CPU rings 10:43 est31 the higher the state is, the more things you can do 10:44 est31 the lower it is, the less you can do 10:44 est31 for cpu rings its the other way round 10:44 est31 but same principle 10:44 TBC_x I know how it works, I just doubt that this is necessary here 10:45 est31 TBC_x, propose something, make a pr and we talk again 10:45 est31 best idea would be to make many small prs 10:45 kilbith so guys you don't see some brief freezes in-game ? 10:45 est31 so that the changes are easier to grasp 10:46 TBC_x wtf, that hand looks really creepy 10:47 est31 hand? 10:48 TBC_x the multicraft one 10:49 kilbith that was not a part of my question 10:49 kilbith just a mistake 10:50 kilbith :/ i see some little freezes since latest RBA shaders commits 10:50 kilbith especially observable when i move the camera 12:21 TBC_x hmm... I'm not completely sure, but there may be used m_env without lock 13:04 TBC_x well... nvm, it's allright 13:42 Zeitgeist_ any opinions on the issue related to -DENABLE_SPATIAL in this pull request https://github.com/gentoo/gentoo/pull/28? 13:43 Zeitgeist_ i am the package maintainer of minetest for gentoo 13:43 Zeitgeist_ and i am not sure if it would break peoples worlds if i enable it unconditionally 13:45 Zeitgeist_ or if its just an optional thing that can be enabled if wished 14:33 Krock Zeitgeist_, it's an optional library and shouldn't change much in the gameplay 14:34 Zeitgeist_ okay i made it optionally avaible as a useflag 14:48 exio4 TBC_x: why should mutable state shared at all, btw? :p 14:48 exio4 (joking a bit.. 14:48 exio4 ) 15:56 est31 Zeitgeist_, Krock is right, spatial is completely optional. For some additional info on the param: I don't know what users use it for right now, but the use case for adding it was to cope with > 30k protected areas on a server with 30 players, from a speed perspective, and even then I have done a second optimisation which should take care of a large part: a simple cache. If spatial is not used, a slower manual implementation is chosen, but the 15:56 est31 speed difference only shows up for very large servers, and if the cache is disabled. Spatial is to make modders not having to care how often they call the areastore functions, or for how irregular positions (which is where the cache fails). 15:57 est31 its fine to have it as optional param 15:57 Zeitgeist_ okay, thank you for the clarification 16:22 rubenwardy Zeitgeist_, it has a fallback implementing in core so no functionality is broken 16:22 rubenwardy it's just faster with that library 16:26 Donillo who is Zefram? 16:26 Donillo technic mod lacks resistance definition for lead blocks in technic/machines/HV/nuclear_reactor.lua 17:05 T4im Donillo: there's #minetest-technic for technic mod related things 17:24 cimbakahn Does anyone know what's happening with stormchaser3000's server? I am getting "Connection timed out" error. 17:24 Calinou cimbakahn, #minetest 17:25 Calinou the server is probably down 17:25 cimbakahn I am not the only one. Others can't get logged in there either. stormchaser3000.noip.me 30002