Time Nick Message 01:21 paramat now pushing 539 to game 01:34 paramat complete. a long overdue job, many were incorrect 02:16 paramat will push this fix for mgv6 soon https://github.com/minetest/minetest/pull/2799 'Mgv6: Don't create air gap in tundra at y = 48 in custom high terrain' 02:21 nolsen I'm not sure if my computer decided to be retarded, but h and / isn't working on minetest. 02:22 est31 pushing a small list-ring fix 02:22 est31 listring[] gives an error, but still works 02:22 est31 I'll remove the error message 02:49 hmmmm est31: I'll have a look at the wiki page you set up 02:49 hmmmm paramat: adds 0.6 seconds to generation time??? 02:49 VanessaE milliseconds. 02:49 hmmmm that's a bit more than i expected... what kind of hardware do you have and are you compiling as release? also I'm just curious, how much of a difference do noisemaps have? 02:50 hmmmm you said that point polling is indeed faster than using a noise map in this case, i would have guessed they'd be around the same in terms of execution speed 02:51 paramat hi 02:53 paramat 0.6 ms. hardware is intel core i5 4670T 2.3GHz x 4. i didn't profile the noisemap method 02:54 paramat point polling adds 0.6ms compared to no biome blend 02:56 hmmmm oh... 02:56 paramat seems well worth it 02:56 hmmmm well try both! 02:56 paramat that's 2 noises with 2 octaves each 02:56 hmmmm what if I was wrong and there's a huge improvement with noise maps 02:56 paramat oh man, i would have to recode the noisemap method 02:57 hmmmm ? 02:57 hmmmm takes 5 seconds 02:57 paramat okay 02:57 hmmmm try in mapgen v5 alone 02:57 hmmmm lines 340, 341, just comment out NoisePerlin2D(... 02:57 paramat after i push 2799, which i'll do now 02:58 hmmmm and add in noise_biome_heat_blend->result[i] and so on 02:58 hmmmm what the heck 02:58 hmmmm oh that's ugly 02:58 hmmmm yeah looks good to me 02:58 paramat maybe someday i'll manage to fix y = 47 bugs in mgv6, difficult though 02:59 hmmmm :S 02:59 hmmmm i should probably do some minetest coding tonight 03:00 paramat and yes trying noisemaps again is not much work =) 03:00 hmmmm anyway yeah. so to convert to noisemaps it's really as simple as commenting out a function call, adding an array access, then commenting out lines 79 and 80, replacing them with noise_heat_blend = ... 03:01 hmmmm and then adding Noise *noise_biome_heat_blend; and so on in the .h 03:01 hmmmm that's 3 modifications 03:01 paramat yes 03:01 hmmmm how did you profile it the last time btw? 03:02 paramat using the profiler in us mode, the code is already there commented out, just timed the loop that adds in the blend variation 03:04 paramat est just pushed something to an unrelated file, is it essential for me to rebase? my pull is 1 commit behind 03:04 hmmmm not unless you're pushing it to upstream 03:05 paramat ah i am, okay will do 03:06 paramat oh maybe it's not upstream actually since it's modifying what was before latest commit 03:06 hmmmm git commit -a --amend 03:08 paramat thanks 03:16 paramat now pushing 2799 03:20 paramat complete 03:49 paramat hmmmm interestingly perlinmap method is 200us, 3 times faster than point polling, unless noise object creation in the ctor takes time, i couldn't time that 03:50 paramat i timed the creation of the blend perlinmaps and the loop that adds those values to heat/humidity maps 03:51 paramat but of course the ctor doesn't run every chunk 03:51 paramat back to perlinmaps then =) 03:58 paramat and all these times are with release build 04:10 nolsen imgcru.sh/foo1234 04:10 nolsen erm 04:10 nolsen http://i.imgur.com/FAu3Ifj.png 04:10 nolsen Mapgen error? 04:10 nolsen Using the moretrees mod 04:11 nolsen I'm not sure if the floating tree been tampered or not. 04:17 nolsen I could try looking for more trees 04:17 est31 nolsen, can you stop crossposting 04:17 nolsen ? 04:18 est31 you keep typing the same stuff in both minetest channels 04:18 nolsen well not the same people is in here. 04:18 est31 some people dont like these practices 04:19 paramat weird, consult Vanessa, in the other channel first in case it's a mod thing 04:19 paramat back later to push noise blend 04:19 nolsen <VanessaE> nolsen: interesting. better bring this up with the mapgen guys. 04:19 nolsen now I'm talking to a ghost 04:20 * nolsen calls the ghostbusters 04:20 nolsen jk 04:39 hmmmm whaa 04:39 hmmmm paramat: okay I stand corrected 04:39 hmmmm paramat: sorry I made you change it in the first place 05:20 paramat no problem, after all the correct thing to do was to time both methods 05:29 paramat nolsen your screenshot looks distorted, do you have shaders and water waves enabled? 05:31 paramat my guess is something to do with fancy leaves intersecting water 05:36 paramat if it's water waves causing this RealBadAngel is the dev to talk to 05:48 paramat pushing https://github.com/minetest/minetest/pull/2764 'Biome API: Noise defined biome blend' very soon 05:52 paramat nore i pushed a commit to game that had 1 approval from sfan5, hope this is acceptable, because it is mapgen stuff and not a new feature i considered my own vote counts, see discussion in this channel 05:52 nore paramat: for is_ground_content? That's fine :) 05:53 paramat okay cool 05:53 paramat yes that commit 05:54 paramat let me know if any of the settings seem wrong 05:54 nore wait, did you add is_ground_content to trees? 05:55 paramat nope, they are false for a good reason apparently 05:55 paramat so i left them that way 05:55 nore yes, that's normal :) 05:55 nore it is to avoid caves carving into them 05:56 paramat yeah 05:57 nore ok, just checked the commit, it is perfectly fine ;) 05:58 paramat good 05:59 paramat i'm working on savanna so will be requesting new nodes soon, acacia tree stuff and golden-brown dry grass and long grasses 06:00 paramat some rainforest will be sunk into shallow water to create a few swamps 06:02 paramat okay now pushing 2764 06:11 paramat complete! =) 07:08 kilbith heyo, does someone can look at #2798 please ? 07:08 ShadowBot https://github.com/minetest/minetest/issues/2798 -- Fix bugs in mainmenu by kilbith 07:09 kilbith ^ sfan5, nore, someone... 09:27 lblot Hi all :) 12:45 crazyR_ i think the settings api need get_number and set_number. 12:45 crazyR_ would save us messing around having to wrap it in tonumber() 13:12 nrzkt crazyR_: having a wrapper is a good idea. 13:56 nolsen <paramat> nolsen your screenshot looks distorted, do you have shaders and water waves enabled? 13:56 nolsen yes. 13:56 nolsen I have the maximum quality settings. 14:43 kilbith nrzkt, does the `dev-0.5` branch can be renamed to `outdated-0.5` ? some server owners use it thinking this is a next-gen branch... 15:27 nrzkt kilbith: why not 15:38 nolsen https://imgur.com/eLs2qyF,As1nPG1,2RgLlxv,LI9yCVa,3ZIf6dL,FjseOT9,Vjpaosf,SMgHhTc#0 15:38 nolsen More info is in #minetest. 15:47 est31 for 0.4.13, due in two months, we need a release manager 15:47 est31 who volunteers :) 15:48 kilbith you ofc 15:48 est31 hah 15:48 est31 release manager dunno if thats fun 15:51 kilbith who else... 15:52 kilbith you're the latest battery of MT after all 15:53 Calinou what is the job of a release manager? 15:53 est31 good question 15:53 Calinou make news posts, edit changelog? 15:53 Calinou do the PR work? 15:54 kilbith keep track the status of things, assess if they are stable enough or not 15:54 kilbith implying a minimum of technical insight 15:54 VanessaE and pull your hair out because there's always some little thing that goes wrong in a release :) 15:55 kahrl make sure all the people who can do builds will be around when it is time to release 15:57 est31 find out who those people are 15:58 VanessaE BlockMen usually isn't it? 15:58 kilbith for MinGW builds 15:59 kilbith hmm, MSVC instead 16:00 rubenwardy Hi, I'd like to work on a mtpm (Minetest Package Manager). It'll be inspired by builtin/modmgr and written in Lua. Should it keep a local list of packages like apt-get does? 16:01 est31 you mean a cache? 16:01 kahrl well don't do it exactly like apt-get, its cache corrupts way too easily :P 16:02 nrzkt i agree to you as a RM est31, you are the more implicated on this release, with paramat 16:04 rubenwardy I think it's a cache. Would it be a good idea? 16:04 est31 I didnt have problems with the cache, but also wonder why its there the first place 16:05 VanessaE I've no objection to a package/mod manager as long as it lets the modder tell it *exactly* where the code comes from (i.e. git) 16:06 rubenwardy ofc 16:06 rubenwardy I will allow mtpm install url 16:07 est31 the only apt get related problems I had were when I caught updating or invalid mirrors (hash not matching downloaded package) 16:07 kahrl est31: I ran into something like this a number of times: http://stackoverflow.com/questions/15505775/debian-apt-packages-hash-sum-mismatch 16:08 kahrl on ubuntu though, not debian 16:09 est31 ah I see 16:09 est31 thanks kahrl next time I encounter that bug, I'll use it 16:09 nrzkt having a good connection and a working hard drive fix the problem. Or use a more stable package manager :p 16:09 crazyR_ why dont we just use luaroacks.. luarock install minetest- ??? 16:09 nrzkt apt-get clean ; apt-get update 16:09 nrzkt and then go 16:10 nrzkt or rm -f /var/cache/apt/archives/*.deb 16:11 est31 perhaps we can fork luarocks 16:11 est31 minetestrocks! 16:13 crazyR_ it just seems a bad idea to re-invent the wheel. lua rocks is brilliant and also allows url install. 16:13 rubenwardy Could you run it from within minetest builtin? 16:13 est31 from looking at one single package (yes really representative), I conclude luarocks seems to be mainly around extending lua by adding c programs 16:13 est31 this again violates mod security principles though 16:13 rubenwardy would it be able to support packages with no metadata? 16:14 rubenwardy ie package.conf 16:15 nrzkt kilbith: renamed to deprecated_dev-0.5 16:15 kilbith thanks 16:15 kahrl when you install a package and then remove it, will luarocks remove the dependencies that were also installed? 16:16 crazyR_ i dont think it will by default. unless ive missed something. but i know you can request a list of dependencies for a paticaular package 16:16 kilbith nrzkt, why have you closed #2411 ?? 16:16 ShadowBot https://github.com/minetest/minetest/issues/2411 -- Colored chat for FreeType builds by nerzhul 16:16 kahrl well, I think that's bad then 16:16 VanessaE aw shit, RBA is here 16:16 est31 I guess because he closed the branch 16:16 nrzkt wtf 16:16 nrzkt i don't close it 16:17 est31 #2447 got closed too 16:17 ShadowBot https://github.com/minetest/minetest/issues/2447 -- Add remote inventory inspection feature by est31 16:17 RealBadAngel awww, whos next to bite? 16:17 nrzkt i only removed dev-0.5 to push deprecated_dev-0.5 16:17 nrzkt oh i see 16:17 kilbith nrzkt, 2418 too 16:17 RealBadAngel howdy folks 16:17 nrzkt the PR was based on dev-0.5 branch... renaming it close the PR... 16:17 VanessaE hi 16:17 est31 heyho 16:18 RealBadAngel VanessaE, est31 minimap is doing great 16:18 nrzkt hi 16:18 VanessaE good deal 16:18 est31 fine 16:18 nrzkt gg to github which automaticly close a PR when the branch isn't existant... 16:18 RealBadAngel i managed to split the task between two threads 16:18 RealBadAngel so its blazing fast 16:18 est31 perhaps we should make a list for later reference 16:19 nrzkt RealBadAngel: great :) 16:19 VanessaE RealBadAngel: link to sources? 16:19 RealBadAngel i tried minimaps >512px already 16:19 RealBadAngel only slowdown is the limit of blocks stored in mem 16:19 nrzkt kilbith: cannot reopen because of the deleted branch.... I note list of closed PR to rebase: 2418 2447 2411 16:19 nrzkt do you see other PR (i don't have notification for PR closed by this deletion) 16:19 kilbith erf :( 16:20 kilbith no only those PRs 16:20 nrzkt i can push the branch to re-open this PR... 16:20 RealBadAngel atm i do store minimap mapblocs 16:20 RealBadAngel which are gettin updated only when mesh has changed 16:21 nrzkt but i think we should rebase those PR to master, because the divergence is too important 16:22 RealBadAngel which reminds me, we need to split highlight, cracks meshes gen from mapblock mesh updates 16:22 RealBadAngel each step of digginin is causing atm new mapblock mesh 16:22 RealBadAngel its a bit sick 16:23 RealBadAngel also, we cannot display properly cracks on each kind of surface 16:23 RealBadAngel my opinion is to drop the cracks at all 16:23 est31 nrzkt, you can probably re-add your PRs like I did with #2801 16:23 ShadowBot https://github.com/minetest/minetest/issues/2801 -- Add remote inventory inspection feature by est31 16:23 VanessaE no, don't drop the cracks 16:23 RealBadAngel and replace them with particels only 16:24 RealBadAngel VanessaE, its not MT 16:24 est31 go to the originating branch, select "create pull request" 16:24 RealBadAngel we dont have to suffer its early sins 16:24 VanessaE cracks show progress in ways that no other method will do 16:25 kilbith agree with VE 16:25 RealBadAngel particles can do the same 16:25 VanessaE besides, if it's causing a whole mesh regen event, seems like a simple solution: overlay a second mesh over the target node and put the cracks *there* instead. 16:25 RealBadAngel from a few, to lotsa of them 16:25 VanessaE copy the target block's mesh to the secondary layer 16:25 VanessaE then animate the cracks over that node-sized mesh 16:26 RealBadAngel mapblock_mesh.cpp isnt a piece of cake 16:26 RealBadAngel its a whole bunch of over engineering 16:27 RealBadAngel we need to separate things 16:32 est31 sounds reasonable, file is 1.4k lines 16:41 est31 nobody stepped up, seems I am the release manager 16:42 est31 how does the date August 10 2015 fit? 16:43 VanessaE seems fine to me. 16:47 est31 ok I guess sfan5 or some other irc op can change the topic then. people will notice I hope. 16:50 kahrl est31: you should be able to (not sure), /msg ChanServ topic #minetest-dev bla 16:51 est31 ~topic 16:51 ShadowBot est31: Minetest core development and maintenance. Chit-chat goes to #minetest. Consider this instead of /msg celeron55. http://irc.minetest.ru/minetest-dev/ http://dev.minetest.net/ 16:54 est31 works 17:09 rubenwardy > not unix time 17:09 rubenwardy XD 17:43 Krock ~ping 17:43 ShadowBot pong 17:45 Krock Doens't look like the WebClient from C# likes the new forum certificate 19:14 IceCreamRenoFail hello, got a few questions about android 19:15 IceCreamRenoFail Where would I look into adding gamepad support? I noticed when I press some buttons i get unexpected code errors 19:40 IceCreamRenoFail is anyone active on here? 19:41 VanessaE at the moment, no, but hang around. 19:41 VanessaE sometimes it's dead, sometimes it's almost too busy to follow. 19:41 IceCreamRenoFail ah alright 20:01 giacomo986 hello everyone, i would like to implement portals in minetest. I already did this with Irrlicht https://www.youtube.com/watch?v=lG3q7mjLg6E . To implement this idea in minetest I need to change parts of minetest core: the scene is rendered 1 time every portal and nodes have to draw on stencil buffers. 20:04 giacomo986 i'm reading minetest code, but i'm not used to programs of this size 21:35 est31 gonna push this: https://github.com/minetest/minetest/pull/2804 21:45 est31 pushed 21:46 est31 and bye again 21:56 paramat hmmmm, looking at l_mapgen.cpp i noticed the mapgen objects heatmap and humiditymap are only enabled for mgv7, they need to be work with any mapgen using the biome api. not sure i know how to implement this properly but i'll look into it 22:07 paramat ..if i follow the example of heightmap and biomemap i might get somewhere