Time Nick Message 02:59 paramat updated game#1541 02:59 ShadowBot https://github.com/minetest/minetest_game/issues/1541 -- Textures: Darker, richer dirt texture by paramat 03:02 VanessaE eck 03:02 VanessaE no way 03:02 VanessaE that dirt is way too dark 03:02 VanessaE it looks almost wet 04:28 paramat yeah dirt is one of those difficult textures, like grass. i'm wasting time maintaining a branch and PR when more experimentation is needed, i'll open an issue instead 04:30 paramat sofar out of interest what's your opinion of game#1541 ? i'm about to close it 04:30 ShadowBot https://github.com/minetest/minetest_game/issues/1541 -- Textures: Darker, richer dirt texture by paramat 04:33 paramat btw i updated game#1587 04:33 ShadowBot https://github.com/minetest/minetest_game/issues/1587 -- Biomes: New surface node for rainforest by paramat 05:01 paramat what do people think about adding krock as an mtgame dev? since he's active, has contributed quite a lot and has been around a long time 05:01 paramat mtgame is understaffed 05:07 sofar paramat: I kind of like it 05:08 sofar it's fairly drastical though, but "weh weh weh" people hate change 05:09 sofar tbh I think we should give up on doing textures one by one and just have a small team redo every single texture at once 05:09 sofar 2-3 people tops 05:09 sofar and no external blocking votes 05:17 paramat i think that would be good for a texture pack, but not for default textures, as contributors will want to submit their own textures for nodes, and if the texture is good and consistent i would disagree with insisting it is redone by someone else 05:19 paramat people generally like the default textures we have, and they are good, there would be a lot of resistence to trashing the lot 05:20 sofar no, you'd just ask the team to add new textures 05:20 sofar and yeah, nothing bad about many textures 05:21 sofar but, if you redo the style consistently, really, you have to throw it all out 05:21 sofar no exception 05:22 paramat my opinion of default textures made by various authors is that it is inevitably less consistent, but needed. texture packs can be (and are already) for small-team / single-artist consistent textures 05:22 sofar seriously if you'd start to use mcresconvert and use some even half-assed texture pack for MC converted to MT, you just don't want to go back 05:22 sofar it's so more pleasing and consistent 05:23 sofar if you want to keep things ugly for old times sake, that's just a really bad proposition 05:23 sofar I could care less about my code getting replaced with something better 05:24 paramat i don't think default textures are that bad, i look at MT texture packs and usually i end up deciding that the default pack is the best we have 05:24 sofar heck a large portion of textures is shamefully ripped from an actual artist 05:24 sofar did you try pixelperfection? 05:24 sofar isabella II? 05:24 sofar hell even the default MC textures look like an eye opener in MT 05:25 paramat our textures are not ugly 05:25 sofar oh, yes they are 05:25 sofar they're absolutely mediocre 05:25 sofar they're inconsistent 05:25 sofar they're not color coordinated 05:25 sofar the contrast level is all over the place 05:25 sofar and so is the saturation level 05:27 sofar show it to an art minor student and they run away crying 05:28 sofar anyway 05:28 paramat hmm, it could be a new texture pack, once complete we could try it and decide to use it as default if it's generally better. then perhaps have the old textures as a shipped alternative pack 05:29 paramat i think our textures are better than MCs even if less consistent 05:29 sofar if only I could get some art students to do a project for school 05:30 paramat well, it's a good idea, we need a variety of texture packs, this could be an official pack 05:31 paramat i'm sure some would maintain the classic textures as an alternative official pack 05:32 sofar find . -name '*.png' -exec cp {} ~/.minetest/textures/minetest-classic/ \; 05:32 sofar there, maintained 05:36 sofar is someone maintaining gambits textures? 05:36 sofar hmm kilbith was for a while 05:43 paramat lots of good textures there 05:44 sofar the whole pack is good 05:44 sofar I admit a few are odd ones out 05:44 sofar but as a whole it's coherent, very 05:44 paramat there are a few i can't stand 05:48 sofar this is why I'm working on something else atm 05:49 sofar can't make any radical progress without throwing out a dozen babies with bathwater 05:58 paramat we do need alternative subgames, and when made by 1 or a few people these can develop many times faster than mtgame does 05:58 paramat free from the PR/review process 05:59 paramat fewer authors means more artistic vision 06:11 paramat i somewhat feel that should be the subgame focus, mtgame work is difficult, being an artistic project with too many authors 06:12 sofar the project I'm working on now sits at 285 commits and 7500 lines of lua code of which barely 1000 were borrowed I think 06:13 sofar not to mention completely new sound and textures 06:14 paramat excellent work 06:14 paramat a variety of subgames has been requested for a long time 06:14 paramat ideally i should make one too 06:15 sofar this is technically not even a subgame project, lol 06:49 paramat #5316 06:49 ShadowBot https://github.com/minetest/minetest/issues/5316 -- Biomes: Add heat and humidity gradient parameters by paramat 07:57 celeron55 i agree that textures should be done by a team that actually knows how to do consistent artwork; is there any chance of that actually happening though, outside of just picking an independent texture pack and using it? 08:00 celeron55 i don't really care though 08:37 kilbith I'm unable to cross-compile MT with GCC-MinGW (arch) : https://hastebin.com/abekewixiq.txt 08:37 kilbith using the buildbot script 09:02 kilbith I really don't understand, yet I have set `export CC=i686-w64-mingw32-gcc` (same for g++) 09:03 kilbith sfan5: any idea ? 09:13 sofar I'm not seeing my minetest ever touch a remote media server O_o 09:17 VanessaE curl support? 09:18 sofar hmm 09:18 sofar oh look there's 2 http download classes 09:19 sofar I put a debug print in the http download setup 09:19 sofar and it only ever downloads the master server list 09:19 sofar no media at all, ever 09:19 VanessaE weird. 09:19 VanessaE clear the cache and try it on my server ? 09:20 VanessaE I'm 100% certain it works there :) 09:20 sofar cleared cache, nothing 09:21 VanessaE huh. 09:21 sofar wait, no new cache made? 09:21 VanessaE ? 09:22 VanessaE I know it sounds obvious, but you're looking in the right place for it I assume? 09:22 sofar I wiped ~/.minetest/cache 09:22 VanessaE it's not there anymore. 09:22 sofar it doesn't get recreated 09:22 VanessaE it's in ~/.cache/minetest/ 09:23 sofar *wipes* 09:23 sofar 500mb lmao 09:23 VanessaE you've been getting around : 09:23 VanessaE :) 09:23 sofar AHHH there we go 09:23 VanessaE I don't remember when that dir was moved but at least that worked :) 09:24 sofar put 60mb of music and wanted to make sure it loads reasonably ok 09:24 sofar (not) 09:24 VanessaE hah! 09:24 VanessaE 10 minute download easy 09:24 sofar 2-3 minutes I estimate 09:24 VanessaE (it takes about 5mins for Dreambuilder's media, which is I think around 30 MB) 09:25 sofar done 09:25 VanessaE sweet 09:25 sofar about 2 09:25 VanessaE guess that works better now :) 09:25 sofar not bad for 60mb of fantastic music 09:25 VanessaE (shows you how long it's been since I last cleared my cache) 09:25 sofar this subgame will rock with this music 09:25 sofar anyway, I so, totally, need to get to bed now 09:25 sofar :) 09:25 * sofar vanishes 09:26 * VanessaE throws flour on sofar 09:26 VanessaE there, now you're visible again. 10:53 nerzhul read(L, -2) ? :) 11:27 glut32 Hi, I would like to compile minetest on MacOSX. I installed Xcode and CMake. And I would like to know wich compiler do you use for the MacOSX compilation with CMake? is it "Xcode" or "Unix Makefile"? 11:28 sfan5 if you want to compile it using GNU "make" from the command like choose makefile 11:28 sfan5 if you'd like xcode then choose that 12:22 nerzhul merging #5132 12:22 ShadowBot https://github.com/minetest/minetest/issues/5132 -- Bug in nodeboxes and collision check. And ghost slabs again, in version 0.4.15 12:22 nerzhul err #5312 12:22 ShadowBot https://github.com/minetest/minetest/issues/5312 -- Fixed spelling mistakes by Bond-009 13:31 glut32 thank you sfan5 ! 13:33 glut32 I want to compile minetest on mac os x. I used "brew install ..." and I launch cmake (compiler = Xcode). And I have this error : "Could not find X11" when I click on configure on CMake. Does someone has a solution to fix this it? 13:38 sfan5 it shouldn't be looking for x11 on mac 13:44 glut32 could you send me your "CMakeCache.txt" file ? 13:45 glut32 I think that a lot path will be the same cuz we use "brew install"? 13:45 glut32 this is my email adress : mr.glut32@gmail.com 13:45 sfan5 i don't have a mac 14:09 glut32 ok. no problem 14:25 glut32 I read the CMakeLists.txt file. You can read this : 14:25 glut32 if(WIN32) 14:25 glut32 ... 14:25 glut32 else() 14:26 glut32 # Unix probably 14:26 glut32 if(BUILD_CLIENT) 14:26 glut32 find_package(X11 REQUIRED) 14:26 glut32 ... 14:27 glut32 endif() 14:27 glut32 It's strange 14:42 celeron55 none of the core developers have macs, you're pretty much on your own 14:42 celeron55 remember to make a PR if something needs to be fixed 14:46 sfan5 this searching for all the deps sucks anyway 14:46 sfan5 it is only needed for static linking 14:46 sfan5 the whole section from if(BUILD_CLIENT) to endif(BUILD_CLIENT) can just be omitted 14:47 sfan5 if responsibility of specifying deps in case of static linking is moved to whoever is compiling mt 14:48 sfan5 glut32: you can comment out the find_package for X11, Xxf86vm, jpeg, bzip2 & png 15:18 glut32 thank you again sfan! the project compiles on xcode but doesnt work. porting::share_path is not valid on mac os x. I'm investigating this now 15:18 sofar glut32: congratulations, you are now the mac port maintainer 15:20 glut32 ha ha 15:21 glut32 * porting::path_share 15:22 octacian Seeing as I'm implementing both a workbench and a nametag to rename items, I'm wondering if I should keep the ability to name nametags with their own formspecs show on_use. What do you guys thing? 15:38 sofar octacian: what's the use of the nametag item if the workbench does all the complex work? 15:38 octacian Yeah, that's what I decided. 15:39 octacian Right now I'm just fighting with getting the workbench to show the name of the item when I move it from the output slot back to the input slot. 15:39 sofar you'd still want a consumable though 15:39 octacian I'm not sure how I can get the description of an item with on_metadata_inventory_move 15:40 octacian Apparently local stack = inv:get_stack(from_list, from_index) doesn't work 15:40 octacian But, up in allow_metadata_inventory_move, inv:get_stack(from_list, from_index) does work 15:40 octacian ..which is a bit confusing. 15:45 octacian All I have left though, is to fix that issue and clean up some code. 15:52 octacian sofar: the nametag is the consumable. One is used with each renaming. 15:52 halt_ Hey anyone know what a PC would need to run a minetest server? 15:53 octacian halt_: it'd be better to ask in #minetest 15:53 halt_ K 16:11 nerzhul halt_: i suggest to have a bi xeon x6 + 64GB ram 16:11 nerzhul MT have memleak and we don't want to fix it, also please use systemd to respawn your service after a OOM kill :p 16:11 nerzhul *joke* 20:11 nerzhul we will need to switch from sha1 to sha256 guys 20:14 est31 why 20:14 est31 we dont use it for any cryptographic purpose 20:15 est31 and in SRP its meaningless which hash is used 20:15 est31 we could use MD5 too 20:15 sfan5 huh srp uses a hash? 20:15 est31 yes 20:15 est31 I mean, weak collision resistance is not what you want from a password hash 20:15 est31 as long as preimage attacks are not feasible, its perfectly usable to use sha1 20:16 est31 (of course, you can say you should use scrypt/bcrypt instead but this will just mean a constant factor) 20:17 est31 we also use sha1 for media files 20:17 est31 I dont see any problems in its use either 20:17 est31 again, we could have used md5 as well and nothing would have went wrong 20:18 nerzhul est31, collision between client files 20:18 nerzhul we use a sha1 to know the unique file 20:18 est31 yes, and? 20:19 nerzhul if two files have the same sha1 your client will just be crazy 20:19 est31 nerzhul: do you know the difference between collision attacks, and second preimage attacks 20:19 sfan5 google's thingy allows everyone to create two colliding pdf files 20:19 sfan5 but i don't see how that is a problem 20:19 sfan5 we don't deal with pdfs 20:19 est31 yeah 20:20 nerzhul it could be extended to other format like images no ? 20:20 est31 yes 20:20 nerzhul and minetest have images, no ? 20:20 est31 yes, so you will see a wrong image, is that any bad? 20:20 sfan5 it's a theoretical attack until someone actually does it 20:20 sfan5 which is estimated to cost like $600k 20:20 nerzhul est31, yes it's bad as it's the wrong image and also could be a rogue image 20:21 est31 nerzhul: but you'd have to download that image first 20:21 est31 I mean minetest already now bases on the concept that the network between client and server is to be trusted 20:21 est31 unless we do encryption ofc 20:22 nerzhul it's wrong to trust client <=> server without encryption 20:22 sfan5 obviously it is 20:23 est31 nerzhul: if the server wants you to send a malicious image, it can send you one already now 20:23 nerzhul base on that concept is wrong and you can send colliding rogue files to clients 20:23 est31 there is nothing preventing it from that 20:23 nerzhul because we use UDP 20:23 nerzhul :) 20:23 est31 server can send all kinds of shit 20:23 sfan5 you don't need a sha1 collision for that 20:23 est31 ^ 20:23 nerzhul sfan5, ofc yes, but why not protect against that ? :) 20:23 sfan5 this is a game not a banking app 20:24 nerzhul sfan5, i think you didn't see banking app code 20:24 sfan5 :D 20:24 est31 :P 20:24 nerzhul the only reason why bank are secure is nobody understand all the infrastructure and code :p 20:25 nerzhul i change the topic but #5088 needs to be merged at a point 20:25 ShadowBot https://github.com/minetest/minetest/issues/5088 -- Client side scripting/modding by nerzhul 20:25 nerzhul i will not lost many time to rebase just for rebase as the API is now stable and secure, only requires a CSM store out of this PR and more callbacks