Minetest logo

IRC log for #minetest-dev, 2015-07-23

| Channels | #minetest-dev index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
00:01 RealBadAngel still the same, all the time pos not ok
00:01 deezl joined #minetest-dev
00:11 RealBadAngel TBC_x, it may look like attachement position is getting corrupted somehow
00:12 RealBadAngel damn idk
00:12 RealBadAngel http://pastie.org/10307053
00:13 RealBadAngel these are coords of all caos that gets lighting updates in my client
00:13 TBC_x do you run it in a debugger?
00:14 RealBadAngel no, im dumping what i need to console
00:14 RealBadAngel notice that there are player coords
00:14 RealBadAngel but there are no similiar coords for attached cao
00:19 TBC_x whoops, It'd be better if it would'nt throw an exception when a server port is in use
00:20 TBC_x now it tries to write 20TB coredump to disk :P
00:29 RealBadAngel trying to get parent of an CAO and get confused
00:29 RealBadAngel error: ‘class ClientActiveObject’ has no member named ‘getParent’
00:30 RealBadAngel https://github.com/minetest/minetest​/blob/master/src/content_cao.h#L125
00:30 RealBadAngel wtf?
00:31 zat joined #minetest-dev
00:35 est31 RealBadAngel, thats GenericCAO
00:35 est31 not ClientActiveObject’
00:36 RealBadAngel class GenericCAO : public ClientActiveObject
00:41 est31 so whats the problem?
00:41 est31 GenericCAO is a ClientActiveObject
00:41 est31 not the other way round
00:41 est31 cast if you are sure you can access the method
00:41 VanessaE RealBadAngel: I wasn't blaming you, I was assigning you ;)
00:44 johnnyjoy I know everyone is busy, but I just wanted to say that I reworked PR 2912. I tried to incorporate ShadowNinja's upsert SQL to combine insert and update, however it was painfully slow compared to the alternative. I also added support for PostgreSQL 9.5(currently in alpha), which should automatically coming insert and update similar to how it's handled by mysql. Once again, thanks for making the world a more fun place.
00:44 RealBadAngel why in content.cao there are includes in the middle of the file?
00:44 est31 RealBadAngel, where?
00:45 RealBadAngel https://github.com/minetest/minetest/​blob/master/src/content_cao.cpp#L539
00:45 RealBadAngel line 329 too
00:45 est31 thats ... bad
00:45 est31 includes always at top
00:45 est31 perhaps this should be fixed.
00:46 RealBadAngel pretty confusing
00:50 TBC_x if C++ could throw rock
00:50 TBC_x the networking code would have mountains
00:52 TBC_x or even planets
00:52 TBC_x of rocks
00:58 OldCoder joined #minetest-dev
01:13 VanessaE est31: I see you and Tesseract have been chewing on that areastore code.  Is it safe to re-add the patch (I forgot to re-apply it after the last updates)?
01:13 est31 VanessaE, lemme see
01:14 VanessaE for that matter, whatever patch(es) the areas mod needs, too
01:14 est31 yes, the problem is that I have to get the areas mod up to date
01:15 est31 Tesseract, any ideas, how can I detect existence of a constructor in lua?
01:16 est31 hrmm leme see
01:23 paramat will push these 3 later game#585 game#586 game#587
01:23 ShadowBot https://github.com/minetes​t/minetest_game/issues/585 -- Flowers: Add mushrooms to mgv6 by paramat
01:23 ShadowBot https://github.com/minetes​t/minetest_game/issues/586 -- Stairs: Register acacia wood stair and slab by paramat
01:23 ShadowBot https://github.com/minetes​t/minetest_game/issues/587 -- Default: Add improved mgv5/mgv7 tree schematics by paramat
01:29 RealBadAngel oke doke, killed that damn bug
01:29 RealBadAngel VanessaE, ^
01:29 VanessaE ?
01:31 RealBadAngel fixed the lighting issues of attached entities
01:31 VanessaE ah
01:31 RealBadAngel http://pastie.org/10307112
01:31 RealBadAngel made all the children use lighting of the parent
01:31 est31 paramat, I dont know if everybody will like it
01:32 est31 can you wait with 587?
01:32 RealBadAngel est31, are you ok with code in pastie?
01:33 est31 RealBadAngel, is that a diff?
01:33 est31 can you show me a diff
01:34 paramat don't worry it's just ongoing improvements to preliminary schems, the current tree schems are not good
01:34 RealBadAngel est31, pastie replaces this function: https://github.com/minetest/minetest/​blob/master/src/content_cao.cpp#L986
01:35 est31 so will it make stuff like in kilbith's screenshot, completely changing how trees look like?
01:38 paramat fairly subtle changes
01:39 est31 so only leaves placement?
01:40 paramat erm i'll add description of changes
01:40 RealBadAngel maybe some screenies too?
01:42 paramat oh yes =)
01:45 VanessaE paramat: not directly related to that commit, but I don't see where you're allowing them to spread after mapgen time
01:45 VanessaE plus there's no apparent way to farm them
01:45 paramat as with the initial biome system, improvement will be an ongoing and frequent process, so i need freedom to update frequently and freely, based on feedback froom users
01:46 paramat i mean 'fairly freely' no one commit will be the final form
01:46 paramat tree spread?
01:46 VanessaE mushrooms
01:47 paramat oh
01:47 VanessaE sorry, shoulda been more specific :)
01:48 paramat erm, flower spread was a bad idea anyway, it should have been done with seeds
01:49 VanessaE mushrooms mod in plantlife uses spores
01:49 VanessaE (which can be had by digging dirt)
01:49 paramat mushroom farming could be discussed and added, if really wanted
01:50 paramat don't see it as much needed
01:50 paramat nore and pilzadam just requested mushrooms on mapgen
01:51 VanessaE pilzadam would ;)
01:51 VanessaE (given his name)
01:57 est31 VanessaE, the prs are now in a state to re-apply
01:58 VanessaE links? :)
01:58 VanessaE well to the areas one anyway
01:58 est31 number 13
01:59 RealBadAngel est31, checked that code?
01:59 est31 RealBadAngel, can you make a commit and push it to your local repo on the master branch
01:59 est31 ?
01:59 est31 and then put the link to the commit here?
02:00 RealBadAngel sure
02:01 VanessaE est31: ok, got 'em.
02:03 est31 RealBadAngel, the change is ok except style issues
02:03 RealBadAngel like?
02:04 est31 1. if(li != m_last_light)
02:04 RealBadAngel brackets on ifs?
02:04 est31 that too
02:04 est31 and dont do updateLightNoCheck (light_at_pos);
02:04 est31 method calls without space, if with space
02:05 est31 otherwise, perfect.
02:05 RealBadAngel such things im already changing
02:05 RealBadAngel most of them are from old code, i meant you to check functionality
02:06 est31 so you have fixed them?
02:12 crazyR who runs/maintains http://servers.minetest.net
02:13 est31 sfan 5 does
02:14 RealBadAngel #2949
02:14 ShadowBot https://github.com/minetest/minetest/issues/2949 -- Fix issues with light of attached CAOs by RealBadAngel
02:14 RealBadAngel if i havent messed anything that should be it
02:14 RealBadAngel VanessaE, can you try it?
02:14 VanessaE sure
02:14 RealBadAngel i need to compile it too
02:15 est31 btw its not required to have brackets around if, just if you have them, then they should be on the same line
02:15 crazyR im just wondering, whether there would be any problem in me scraping http://servers.minetest.net/list every ~120 seconds.
02:16 VanessaE crazyR: I thought there was some API access already?
02:16 VanessaE RealBadAngel: does not compile
02:16 VanessaE void GenericCAO::updateLightNoCheck(u8 light_at_pos)
02:16 VanessaE er..
02:16 VanessaE /home/vanessa/Minetest-related/minete​st_core/src/content_cao.cpp:1004:52: error: no ‘void GenericCAO::updateLightNoCheck(irr::u8)’ member function declared in class ‘GenericCAO’
02:16 est31 RealBadAngel, add it to GenericCAO
02:17 crazyR VanessaE im not sure, in my tests ive just grabbed the json data from http://servers.minetest.net/list and parsed it. but i dont want to end up having my servers blocked if its not allowed
02:17 RealBadAngel i knew i messed something ;) hold on
02:18 RealBadAngel void updateLightNoChek(u8 light_at_pos);
02:18 RealBadAngel lmao
02:22 RealBadAngel VanessaE, est31 updated
02:23 est31 why have you added virtual void updateLightNoCheck(u8 light_at_pos){} into ClientActiveObject?
02:23 VanessaE ok
02:24 RealBadAngel because otherwise i couldnt call it
02:24 RealBadAngel children are ClientActiveObjects
02:24 est31 ah
02:24 est31 ok then
02:24 est31 +1, you can push it
02:24 est31 (if VE reports it works)
02:24 VanessaE I'm still setting up to test :P
02:25 RealBadAngel i hope so, it worked for me
02:26 VanessaE works for me.
02:26 VanessaE +1
02:26 paramat okay screenshot and description added to game#587
02:26 ShadowBot https://github.com/minetes​t/minetest_game/issues/587 -- Default: Add improved mgv5/mgv7 tree schematics by paramat
02:26 RealBadAngel paramat, can you +1 my pr too?
02:28 est31 RealBadAngel, its a small fix, only needs one +1
02:28 est31 the second person decides
02:28 est31 at least this is consensus I think
02:29 est31 paramat, the apple tree is now much more "flat", no?
02:29 paramat i trust your tests +1 i'm a bit busy
02:30 paramat yes appears flatter due to being at edge of view, actually the same outer dimensions as before
02:31 est31 are you sure?
02:31 paramat very =)
02:32 paramat it has height variation, that one is the shorter one
02:35 paramat remember don't compare to mgv6 trees, these are for mgv5/v7, different designs
02:36 paramat back later
02:36 paramat left #minetest-dev
02:49 VanessaE est31: those datastore patches are staged now.  should go in at tonight's reboot.
02:57 est31 areastore*
02:58 VanessaE er yeah
02:58 VanessaE that :P
03:55 OldCoder joined #minetest-dev
04:07 paramat joined #minetest-dev
04:08 Hunterz joined #minetest-dev
04:09 paramat i'll wait a day for game 587, now pushing game 585 and game 586
04:18 paramat complete!
04:18 VanessaE staged those to my server :)
04:18 VanessaE so if they break, you'll know ;)
04:20 VanessaE bbl
04:20 paramat i will indeed =D
04:43 paramat left #minetest-dev
05:03 hmmmm est31, I looked at 2922
05:03 hmmmm looks great overall
05:04 hmmmm don't think I saw any bugs
05:04 est31 nice
05:07 est31 commented on your comments, some need a reply
05:07 hmmmm we're both replying in real time
05:08 est31 is DenyAccessVerCompliant ok too?
05:09 hmmmm yea
05:09 hmmmm 'cause it confused me before you explained it
05:10 hmmmm code written for other people needs to be realllly obvious
05:10 hmmmm @RBA
05:11 est31 yea agreed, I wont be around forever to explain my code, it has to explain itself :)
05:32 hmmmm alright
05:32 hmmmm since that field is a u8 and not strictly a bool, that reconnect field can get changed later on
05:32 hmmmm looks good to me
05:34 est31 can I push 2922?
05:34 hmmmm yeah
05:38 est31 pushed
05:43 Hunterz joined #minetest-dev
06:08 Taoki joined #minetest-dev
06:28 OldCoder joined #minetest-dev
06:44 GunshipPenguin joined #minetest-dev
06:45 GunshipPenguin left #minetest-dev
06:45 nore joined #minetest-dev
06:46 Darcidride joined #minetest-dev
07:08 Krock joined #minetest-dev
07:18 FR^2 joined #minetest-dev
07:46 jin_xi joined #minetest-dev
07:53 kilbith joined #minetest-dev
08:00 Yepoleb_ joined #minetest-dev
08:26 Taoki_1 joined #minetest-dev
08:33 julienrat joined #minetest-dev
08:33 julienrat left #minetest-dev
08:51 leat1 joined #minetest-dev
08:58 TBC_x joined #minetest-dev
09:00 leat1 joined #minetest-dev
09:11 TBC_x joined #minetest-dev
09:16 yh1986 joined #minetest-dev
09:17 yh1986 hi, why sky don't work on android
09:17 yh1986 no sun、moon and stars
10:15 Calinou joined #minetest-dev
10:33 RealBadAngel yh1986, what build are you using? you should fill an issue describing your problem
10:33 RealBadAngel https://github.com/minetest/minetest/issues
10:33 RealBadAngel so devs can solve that
10:44 est31 joined #minetest-dev
10:45 est31 ok, now I see why kwolekr regarded the minimap updates as wasteful.
10:50 blaze joined #minetest-dev
10:54 yh1986 Minetest-0.4.12.11-android this
10:55 yh1986 official version also have the issue
10:56 est31 yh1986, what do you mean sky isnt working?
10:56 est31 can you make a screenshot?
11:00 yh1986 sun、moon、and stars can't render success
11:00 yh1986 Sky::render()
11:01 est31 yh1986, screenshots say more than 1000 words
11:02 yh1986 Sky::render() not work on android
11:05 RealBadAngel est31, what do you mean with "wasteful"?
11:05 est31 RealBadAngel, getMinimapNodes is called every time we add a new MapBlockMesh
11:06 est31 and that one scans the whole map
11:06 kilbith joined #minetest-dev
11:06 est31 this generates needless cpu
11:07 RealBadAngel its not quite like that
11:07 RealBadAngel on mesh update you get 16x16x16 scan
11:07 est31 no
11:07 RealBadAngel until this block doesnt change its not triggered anymore
11:07 est31 we call getMinimapNodes
11:08 RealBadAngel yes, on mesh update
11:08 est31 that has xzy loops between 0 and MAP_BLOCKSIZE
11:08 RealBadAngel yes
11:08 est31 and its called on every mesh updat
11:08 est31 e
11:08 RealBadAngel yes
11:08 yh1986 left #minetest-dev
11:09 est31 so there is no 16x16x16 scan at all
11:09 RealBadAngel which happens at worst a few times a second, if you dig in that block
11:09 RealBadAngel its pre-scan
11:09 est31 it happens every time something changes in the block
11:09 RealBadAngel yes, but it one block
11:10 est31 ! 512 ^3
11:10 RealBadAngel the one (or its neighbours) youre fiddling with
11:10 est31 !c 512 ^3
11:10 ShadowBot est31: math calc <math expression>
11:10 RealBadAngel 16^3
11:10 est31 !c 512 * 512 *512
11:10 ShadowBot est31: math calc <math expression>
11:10 est31 ~c 512 * 512 *512
11:10 ShadowBot est31: math calc <math expression>
11:10 est31 either way
11:10 RealBadAngel 512^3 != 16^3 :P
11:10 est31 in the worst case, it scans really 512*512*512 nodes
11:10 RealBadAngel who told you that?
11:11 RealBadAngel hmm?
11:11 est31 I see that in the code
11:11 RealBadAngel you can see it wrong then
11:11 RealBadAngel shall i write a book on the idea? ;)
11:11 RealBadAngel if you were right realtime scanning would never be possible
11:11 est31 ah I see
11:11 est31 thats ok
11:12 RealBadAngel whole scan is circa 8,6kk of nodes
11:12 est31 no problem here then
11:12 RealBadAngel ive counted it
11:12 RealBadAngel but thx to mapblock mesh usage, most of the scan is done once
11:12 RealBadAngel or just on update
11:13 RealBadAngel so in general full scan means 8,6kk / mapblock size
11:13 RealBadAngel and thats doable realtime
11:14 RealBadAngel and even most of the fps drop caused is not by scan itself but texture compositioning each frame
11:15 RealBadAngel i could make it faster by further caching areas but....
11:15 RealBadAngel we dont have one surface level
11:15 RealBadAngel we are supposed to get multilevel worlds
11:15 RealBadAngel minimap has to stay in touch where the player is
11:16 RealBadAngel fly a bit above, youre in next realm, with another surface
11:16 julienrat joined #minetest-dev
11:16 RealBadAngel or you just have met some structures in the sky
11:17 RealBadAngel current code can handle that
11:19 Darcidride joined #minetest-dev
11:21 leat1 joined #minetest-dev
11:21 RealBadAngel est31, also 512 is not max scan dimension, its the max resolution of an image
11:22 RealBadAngel hmmm have messed it up
11:22 RealBadAngel max scan (in term of nodes is 256^3 with an option to reduce it by half)
11:23 est31 getMinimapPixel should work on MinimapMapblocks instead of pixels
11:23 est31 you know, retrieve a whole MinimapMapblock
11:24 est31 then cache that one for the later search
11:24 est31 spares you repeated calls to m_blocks_cache
11:24 RealBadAngel thats a real chalenge
11:24 RealBadAngel thats a place we propably waste most cpu power
11:25 RealBadAngel blocks are well, blocks
11:25 RealBadAngel scan needs to check it node by node
11:25 est31 thats perfectly ok
11:25 RealBadAngel we are wasting time here to conver node pos to block pos
11:26 RealBadAngel thats not the problem on Y
11:26 RealBadAngel x and z are the most consuming
11:26 est31 agreed, that can be spared
11:27 RealBadAngel propably precalculatint starting offset, number of full blocks and end offset would help
11:27 RealBadAngel at least in making highly optimized loops
11:28 RealBadAngel atm i cant see a simple and effective solution in mind yet
11:28 est31 also we should cache the scan result
11:28 est31 unless there is actual change
11:28 RealBadAngel you do scan cache already
11:29 est31 yes
11:29 kilbith joined #minetest-dev
11:29 est31 but that cache only has mapblocks
11:29 kilbith est31: why 2922 is still opened ?
11:29 RealBadAngel but notice that raising a node up can change minimap look
11:29 RealBadAngel no matter what in cache is
11:29 RealBadAngel you are working on pre-processed data
11:30 RealBadAngel also moving in each direction
11:30 est31 kilbith, closed
11:32 RealBadAngel i dont think that double caching there would be of any help
11:32 kilbith RealBadAngel: have you reviewed #2946 ?
11:32 ShadowBot https://github.com/minetest/minetest/issues/2946 -- Fixed minimap memory leak by t0suj4
11:32 est31 kilbith, we are waiting for TBC_x to update the pr to hmmmm's suggestions
11:32 RealBadAngel kilbith, yes, that code is broken. i already told him so and requested to stop on that code
11:33 RealBadAngel radar mode is broken thx to hmm's changes. TBC_x following this direction is makin it even worse
11:33 kilbith well that's not surprising, he confused the mapblocks caching with the minimap leak
11:34 RealBadAngel i was delayed by lighting issues of wielded items, but ive already fixed that
11:34 RealBadAngel now im on minimap back again
11:34 RealBadAngel and im going to shoot down anybody else who will attempt to break it
11:35 est31 how is radar mode broken?
11:35 RealBadAngel go on the surface
11:35 RealBadAngel you should see kinda night vision one
11:36 RealBadAngel all you can see now is whole green area
11:36 est31 yea
11:36 est31 is that thanks to hmm's changes?
11:36 est31 interesting
11:36 RealBadAngel yes
11:36 RealBadAngel remember asking me to fix the squares in radar mode?
11:36 est31 yes
11:37 RealBadAngel i found a solution and make it look like real night vision googles style
11:37 RealBadAngel they broke it
11:38 est31 ok
11:43 kilbith we may need to nuke the rendering dropdown in the settings tab, as sfan5 suggested
11:43 kilbith we lack of place in that confined space
11:44 kilbith perhaps replacing it by more useful settings like viewing_range or mouse options
11:46 RealBadAngel kilbith, an external irrlicht app that will just do settings is the solution
11:46 RealBadAngel you could have there books on settings then ;)
11:47 kilbith it's maybe too technical for the regular users
11:47 RealBadAngel while keeping main game menus simple and compact
11:47 RealBadAngel for those shaders on/off will do
11:48 kilbith moving plants/water/etc. can be merged in one setting already
11:48 RealBadAngel if you wanna dig deeper, launch minetest-config and have fun
11:48 RealBadAngel within shaders lotsa things will change soon
11:49 RealBadAngel and there will be tons of new settings
11:49 kilbith so we maybe need a dedicated formspec for shaders
11:49 RealBadAngel such small windows we are having now are of no use for me
11:49 kilbith [advanced graphics]
11:50 RealBadAngel we are talking about lotsa sliders, dozens of options
11:50 RealBadAngel forget it
11:50 RealBadAngel current layout is not enough
11:51 RealBadAngel once im done with current issues i plan to get merged REAL advanced effects
11:51 kilbith yeah
11:51 kilbith everything is tight as pickles in a jar
11:51 RealBadAngel like shadows, reflections, refractions, postprocessing and all the fun
11:52 kilbith irrlicht dynamic lighting, with sun as real source
11:52 RealBadAngel we do have working tangent space now
11:52 RealBadAngel thats a basis for everything
11:53 RealBadAngel no more faking it, weird tricks to get even most simple effects working
11:53 RealBadAngel now we can just squeeze GPU to get them all
11:54 * kilbith excites
11:54 RealBadAngel relief mapping and how it looks right now is just a begining
11:54 Hunterz kilbith: I dont know player who love edit game config using text editor :P
11:55 kilbith me neither, this is why we try to integrate them in the mainmenu
11:58 Hunterz oki
12:02 proller joined #minetest-dev
12:10 proller joined #minetest-dev
12:14 FR^2 joined #minetest-dev
12:25 FR^2 joined #minetest-dev
12:26 julienrat joined #minetest-dev
12:29 err404 joined #minetest-dev
12:31 RealBadAngel kilbith, Hunterz im not talkin bout text one. im talkin about standalone settings oriented one only
12:32 Hunterz ok
12:33 RealBadAngel with it you wont be limited by number of options
12:33 RealBadAngel it will be a config app
12:33 RealBadAngel also, there wont be any formspecs
12:34 RealBadAngel only native irrlicht GUI elements
12:34 Hunterz Im only tried resolve own issue :)
12:34 RealBadAngel main menu should be a kiss one
12:35 RealBadAngel we, as the team, and the community are forgetting bout it
12:35 RealBadAngel game should be able to run as click n go
12:35 kilbith yeah, fool-proof
12:36 RealBadAngel fool is a bit too strong word
12:37 RealBadAngel just a newcomer. clicked, played, satisfied == will come back
12:37 kilbith the ultimate referrence is still MC
12:37 RealBadAngel i was playin mc, having a workin setup of my fav mods was pain in the ass
12:38 RealBadAngel and im a coder
12:38 kilbith it's not meant to be modded either
12:38 kilbith but the menu is just perfect
12:38 RealBadAngel no wonder there are how-to's for it all around
12:38 WSDguy2014 joined #minetest-dev
12:38 WSDguy2014 left #minetest-dev
12:38 WSDguy2014 joined #minetest-dev
12:39 RealBadAngel we shall avoid giving a reason to write a how-to
12:39 WSDguy2014 left #minetest-dev
12:39 RealBadAngel we are not the pioneers, we cant afford to be comlicated
12:40 RealBadAngel on the very first problem newcomer will call us shitlike and go back to mc
12:41 RealBadAngel even when its much more complicated there, but well documented
12:42 kilbith because MC in an accomplished game, MT is more a playground for coders
12:42 RealBadAngel it is not a finished project
12:42 RealBadAngel sandbox out of its def will never be
12:43 RealBadAngel you will think one day, heck why im not able to fuck an apple? thats so obvious for me ;)
12:43 Darcidride joined #minetest-dev
12:43 RealBadAngel well, its a sandbox, you are supposed to do whatever you want to ;)
12:45 Hunterz BFU player want play game, not study Doc iho
12:45 Hunterz imho
12:48 Hunterz As server owner there are lots of missing features for me:  server console, privilege groups, chat prefixes depend groups, lua mysql support, multi world support... etc...
12:51 RealBadAngel Hunterz, notice we do start talkin about very game oriented settings
12:52 RealBadAngel that each game would want to have defaulted
12:52 Hunterz sorry
12:52 RealBadAngel thats a quantum leap forward imho
12:53 Megaf joined #minetest-dev
12:54 RealBadAngel Hunterz, ofc you are now defining general features missing
12:54 RealBadAngel sooner or later we will get them all
12:54 RealBadAngel wonder what devs will do then
12:55 RealBadAngel will community fund us at least a weeken on Mallorca? ;)
13:19 SopaXT joined #minetest-dev
13:36 H-H-H joined #minetest-dev
13:38 julienrat left #minetest-dev
14:11 FR^2 joined #minetest-dev
14:15 proller joined #minetest-dev
14:18 Hunterz joined #minetest-dev
14:27 Darcidride joined #minetest-dev
14:31 TBC_x RBA, I would really appreciate if you would stop blaming me that I broke your radar code
14:47 Darcidride joined #minetest-dev
15:04 Darcidride joined #minetest-dev
15:07 Cornernote_ joined #minetest-dev
15:35 MinetestForFun joined #minetest-dev
16:07 Hunterz joined #minetest-dev
16:12 Krock joined #minetest-dev
16:22 Robert_Zenz joined #minetest-dev
16:28 hmmmm joined #minetest-dev
16:38 SopaXT joined #minetest-dev
16:46 TBC_x who was messing with wieldmesh lately?
16:47 TBC_x whoever did it, he did a poor job
16:48 TBC_x the code was not leaking there a couple of commits ago
16:48 kilbith joined #minetest-dev
16:55 kilbith left #minetest-dev
16:56 sloantothebone joined #minetest-dev
17:06 FR^2 joined #minetest-dev
17:25 Player_2 joined #minetest-dev
17:47 err404 joined #minetest-dev
18:09 proller joined #minetest-dev
18:16 err404 joined #minetest-dev
18:17 est31 joined #minetest-dev
18:17 est31 TBC_x, RBA "messed" with it
18:17 est31 he fixed a bug
18:18 est31 that was the original intention
18:18 TBC_x oh, cmon
18:18 est31 but he also added a shader
18:18 TBC_x I've noticed 3D stick
18:18 est31 ?
18:18 TBC_x if that was the shader
18:18 est31 you mean extrusion?
18:19 est31 it was there before that
18:19 TBC_x yeah, that leaky freak
18:19 est31 no, extrusion in this context means creating a 3d shape from a 2d image
18:20 est31 perhaps you have ideas to fix?
18:20 est31 also, which tools do you use to determine the leaks?
18:20 TBC_x -fsanitizer=address -fsanitizer=undefined
18:20 est31 thanks
18:21 TBC_x in gcc
18:21 TBC_x gcc version 5.1.0 (GCC)
18:21 est31 can you make a bisect to be really sure?
18:21 TBC_x define bisect
18:21 est31 git bisect
18:21 TBC_x oh, I don't know that command
18:23 Krock a git GUI can be useful in this case
18:23 TBC_x well... I know that the leak was not there before I rebased to d569c91f
18:23 proller joined #minetest-dev
18:23 TBC_x I am doing final tests on the minimap leak patch
18:24 est31 thats a start
18:24 est31 also I think RealBadAngel blamed hmmmm to break the minimap not you
18:25 TBC_x the game loop needs IMO heavy refactoring
18:25 VanessaE [07-23 07:22] <RealBadAngel> hmmm have messed it up
18:26 TBC_x before I moved on to minetest I was developing entity-component system in C
18:27 TBC_x but i abandoned it because i skipped the first step -> prototyping in interpreted language
18:32 proller joined #minetest-dev
18:32 TBC_x though, radar mode quadruples my drawtime
18:32 TBC_x but i shouldn't jum into conlusions with -O0
18:33 TBC_x s/jum/jump/
18:35 VanessaE weird.
18:35 VanessaE the minimap (any mode) doesn't have a noticeable effect on my fps/drawtime
18:36 VanessaE but then again my machine performs totally sub-par compared to others
18:36 TBC_x what graphics card do you have?
18:36 VanessaE AMD R9 280X.
18:37 TBC_x 00:02.0 VGA compatible controller: Intel Corporation Mobile GM965/GL960 Integrat
18:37 TBC_x ed Graphics Controller (primary) (rev 03)
18:37 TBC_x whoops, line wrapper
18:37 VanessaE ah
18:39 VanessaE maybe it's time to rip the whole rendering pipeline apart and rewrite it :P
18:39 TBC_x got leak in src/util/srp.cpp:810 if anyone wants to take it apart
18:40 VanessaE there's just no legit reason why a single tile floating over the rest of the display should be any slower than, say, adding another isolated node to the scene.
18:40 TBC_x maybe the reason is -O0
18:41 VanessaE I wouldn't think so, really
18:42 TBC_x btw what's up with *NoEx methods?
18:42 VanessaE no idea
18:42 TBC_x shouldn't that get cleaned up?
18:44 TBC_x and i hate MinimapMapblock name
18:44 leat1 joined #minetest-dev
18:44 RealBadAngel lol
18:44 TBC_x i never spell it right the first time
18:45 RealBadAngel me neither ;)
18:45 TBC_x then cut the extra map
18:46 RealBadAngel its derrived from the place where its created
18:47 RealBadAngel i could call it propably "pre-scan data" or something like that, but mapblock mesh and its thread suggested something similiar
18:48 TBC_x MinimapBlock would suffice
18:48 TBC_x or MinimapTile
18:49 RealBadAngel btw, i started to analyze how often its being called
18:49 TBC_x good thing
18:49 TBC_x for me unnecessarily too often
18:49 RealBadAngel and i found out that moving all the toys outta there will be the best idea ever
18:49 TBC_x just wanted to put up this suggestion
18:50 RealBadAngel i figured out why duplicated entries are being deleted from update queue
18:50 TBC_x why?
18:50 RealBadAngel queue is actually flooded with refresh requests
18:51 RealBadAngel loading a block means 7 updates, main + 6ddir
18:51 TBC_x would be better to reduce the amount of updates in the beginning
18:52 TBC_x unless it would break something
18:52 RealBadAngel no can do colonel, addUpdateMeshTaskWithEdge
18:52 RealBadAngel meshes can change if neigbour ones are changing
18:53 RealBadAngel theyre not static
18:53 RealBadAngel also digging triggers the update a few times
18:54 RealBadAngel best solution is to get the toys outa here and run as fast as possible
18:54 TBC_x I think we should find a way to show cracks without touching the mesh
18:54 RealBadAngel i already tried
18:54 TBC_x and?
18:54 RealBadAngel dont even try to do that
18:55 TBC_x what happens?
18:55 RealBadAngel fanatics of the cracks religion will hunt you down
18:55 TBC_x haha
18:55 RealBadAngel i tried to replace cracks with highlighting animations and some particles already
18:56 TBC_x sounds dumb
18:56 Calinou cracks are still useful, we shall keep them
18:56 RealBadAngel see? they started it already ;)
18:56 jin_xi you fool
18:56 RealBadAngel better hide ;)
18:56 TBC_x I had in mind drawing the crack animation over the block being dug
18:57 RealBadAngel no can do
18:57 RealBadAngel it works for whole blocks
18:57 Calinou that'd double the polycount for any mesh
18:57 RealBadAngel but you will fail on plantlike for example
18:58 TBC_x plantlikes don't have their own meshes?
18:58 RealBadAngel plantlike are just 2 planes
18:58 RealBadAngel to make cracks with outer mesh you shall have it double
18:58 TBC_x apparently I haven't wrapped my head around mesh concept
18:59 RealBadAngel a plane before one plant plane, one behin it
18:59 RealBadAngel and the same for another one
18:59 TBC_x does the block highlight box update the mesh?
19:00 RealBadAngel yup, but thats already coded and moved to separate scene node
19:00 RealBadAngel so highlighting needs to call mesh update thread only once and only for single mesh
19:00 TBC_x so the minimap mesh is updated every time i highlight another node?
19:01 RealBadAngel after receiving the mesh it pisses on it
19:01 jin_xi taking a leak
19:01 * jin_xi hides
19:01 err404 joined #minetest-dev
19:02 RealBadAngel TBC_x, its not as bad
19:02 TBC_x depends
19:02 err404 left #minetest-dev
19:02 RealBadAngel digging whole node causes 5-6 mesh updates to the block
19:02 RealBadAngel just put some dstream <<"Here!" into the code and you will see
19:03 TBC_x so dstream
19:03 TBC_x someone suggested errorstream to me
19:03 RealBadAngel errorstream slows it even more, because it writes to the screen
19:03 VanessaE you take away my cracks animation and I take an admin pick to your ass :)
19:04 RealBadAngel not only the console
19:04 TBC_x who was talking about taking away cracks?
19:04 RealBadAngel dstream is safer
19:04 RealBadAngel TBC_x, lemme find a vid on that
19:04 VanessaE I'm just a little behind, TBC :P
19:04 TBC_x vid on what?
19:06 RealBadAngel https://www.youtube.com/watch?v=wjG1bUBFPB0
19:06 TBC_x I suggest to compile at least with -fsanitizer=leak before anyone pushes their changes
19:07 TBC_x oh, that weird animation
19:08 TBC_x I think i saw it somewhere already
19:10 RealBadAngel https://www.youtube.com/watch?v=hGv6k3-2D3s
19:10 RealBadAngel i just reminded myself that
19:10 RealBadAngel that has to be done sooner or later
19:10 paramat joined #minetest-dev
19:11 RealBadAngel but for that we would need VBO running
19:11 RealBadAngel its too costly
19:11 VanessaE imagine that ^^^^ extruded video, but with HDX torches.
19:11 paramat hi nore, sfan5, please can i push game#587 later?
19:11 ShadowBot https://github.com/minetes​t/minetest_game/issues/587 -- Default: Add improved mgv5/mgv7 tree schematics by paramat
19:11 VanessaE talk about mesh abuse :P
19:11 RealBadAngel paramat, +1
19:11 VanessaE paramat: wait, I wanted to say,
19:12 VanessaE paramat: I think the pines would look better if you could round them off a bit, as seen from above
19:12 paramat yes, i can reduce the probability of the corner nodes
19:12 RealBadAngel VanessaE, there are different kind of pines
19:12 RealBadAngel he could follow more trees one too
19:13 VanessaE RealBadAngel: yeah, and no pine I've ever seen has a square appearance from above ;)
19:13 VanessaE paramat: ok.
19:13 TBC_x why mods can't define their own drawtypes anyway?
19:13 RealBadAngel TBC_x, do you actually examined what a drawtype means for the engine?
19:13 paramat it's deliberately a bit squarish and naive to be similar to mgv6 pines
19:14 TBC_x i did not, so I am asking
19:14 RealBadAngel content_mapblock.cpp
19:14 VanessaE paramat: oh wait, I was thinking of the mgv6 pines.
19:14 RealBadAngel and dont ask again if mods can do that ;)
19:15 RealBadAngel only engine level mods - patches to the code are able to do such gimmicks ;)
19:16 paramat well anyway i checked my schematic code, the pine needs a little work anyway, and the corners do need less probability, so i'll update
19:16 RealBadAngel paramat, you can tweak the schematics later on based on feedback
19:16 Taoki[mobile] joined #minetest-dev
19:17 MinetestForFun joined #minetest-dev
19:17 RealBadAngel dont even try to have best possible one with the initial commit
19:21 RealBadAngel TBC_x, btw, https://www.youtube.com/watch?v=2x6w0wspmPo
19:22 RealBadAngel and https://www.youtube.com/watch?v=hubBBzh_190
19:22 RealBadAngel these were faulty before, no tangent space aviable
19:23 VanessaE bah, we have other more important things to work on
19:23 RealBadAngel VanessaE, i know, thats why i do wait with those for next stable
19:23 RealBadAngel after that we will have water with reflections
19:25 RealBadAngel adding new shaders is now dumb easy, tangent space is here
19:25 RealBadAngel i just have to add double buffering and rendering passes
19:28 RealBadAngel mt will be able to use most advanced rendering thingies out there
19:28 VanessaE > wants to use super-advanced rendering
19:28 RealBadAngel nearly out the box
19:28 VanessaE > engine stuck in single-threaded GPU access
19:29 RealBadAngel oh cmon, remember first games for c64 or zx spectrum?
19:29 VanessaE of course.
19:29 RealBadAngel now remind yourself the last titles
19:30 RealBadAngel hardware was not changing
19:30 TBC_x suddendly everything is shit compared to lava
19:30 RealBadAngel TBC_x, btw, that texture is 16px ;)
19:31 TBC_x ikr
19:31 RealBadAngel rest is done with math
19:31 TBC_x when I saw that texture for the first time, it was the most shitty texture in the entire game
19:32 TBC_x apart from grass with dirt in mt 3.x
19:32 RealBadAngel water is even worse, theres no texture for it at all ;)
19:32 TBC_x how many AA passes the lava has? unless it uses something else
19:33 TBC_x I am very unfamiliar with computer graphics unfortunately
19:33 RealBadAngel its single pass
19:33 TBC_x hard to believe it is 16x16
19:34 RealBadAngel adding some bumpmapping makes it look like its more detailed
19:34 RealBadAngel same effect as im using now with relief
19:34 TBC_x try to put arbitrary texture to lava and see how it looks
19:35 TBC_x just 4 fun
19:35 RealBadAngel there are some examples uploaded already
19:35 RealBadAngel hard to say which one is better, 16px or 512px
19:36 RealBadAngel https://www.youtube.com/watch?v=v52f-1YsSAM
19:36 TBC_x this one looks kinda awful
19:37 RealBadAngel https://www.youtube.com/watch?v=3-o1mGWp1A0
19:37 RealBadAngel ^^ 512px
19:40 RealBadAngel https://www.youtube.com/watch?v=vF9FZx2f8Uw
19:40 TBC_x the 16px one looks better IMHO
19:40 RealBadAngel on the last one you can see 2 textures at the same time (upper left corner is simple 16px without bumpmapping)
19:40 RealBadAngel rest is 256px with normalmap
19:57 RealBadAngel anyway, those are for the future, now we have more important things to do
20:07 TBC_x I'll try to add mutex to irr::IReferenceCounted and try to not lock up
20:07 TBC_x that reference counting bugs are bugging me
20:07 kilbith joined #minetest-dev
20:09 TBC_x hmm
20:09 TBC_x or just ignore that indirect leak
20:09 kilbith sorry to say but since latest RBA's commit for shaders on wieldhand, one face is "noised" with lotsa of white-pixels (amost unoticeable) : https://lut.im/BinxM4Kl/rRIBCNQW
20:10 kilbith *almost
20:11 RealBadAngel pixels were here
20:11 kilbith absolutely not
20:11 RealBadAngel wanna bet? ;)
20:11 kilbith i have a good memory
20:11 RealBadAngel shading made it just more visible
20:11 paramat the dots are on the edges of the pixels in the hand texture
20:12 RealBadAngel theyre on edges of a mesh
20:12 kilbith paramat: not on the edges, and they're random
20:12 RealBadAngel theyre not random
20:12 kilbith they cover the whole face, you dont see ?
20:12 RealBadAngel mesh is made out of small cubes
20:13 RealBadAngel cubes are next to each other
20:13 kilbith ah, ok - so whether i'm a fuckin liar or i have hallucination
20:13 kilbith thanks RBA
20:13 RealBadAngel effect is the same killbith
20:13 RealBadAngel but the reason is different
20:13 kilbith do you see the white pixel on the left face ??
20:13 RealBadAngel we need fine tuned mesh for wielded
20:14 paramat vertically random yes, but they run along 5 equally-spaced vertical lines that separate the 6 columns of pixels
20:14 kilbith ...
20:14 kilbith you just don't see it
20:14 kilbith i'm not talking about straight lined pixels
20:15 kilbith there are pixels covering randomly the left face while i walk
20:15 RealBadAngel kilbith, you just see a plain had texture. i can see 32x32 small CUBES with that texture mapped bit by bit on those
20:15 RealBadAngel connection between those cubes makes those artifacts
20:15 RealBadAngel *hand
20:16 paramat okay i'll check in-game, this is shaders off i guess
20:16 RealBadAngel that is not shaders dependent
20:16 RealBadAngel shaders thx to bumpmapping can make it more visible just
20:17 kilbith https://lut.im/VFysYTKr/6eIFQorG
20:17 kilbith you see better ?
20:17 RealBadAngel ofc i can see it
20:17 kilbith they are ramdomly spread while i walk
20:18 kilbith i swear there was not that before
20:18 RealBadAngel irrlicht rendering accuracy
20:18 VanessaE kilbith: http://digitalaudioconcepts.com/vanes​sa/hobbies/minetest/screenshots/rando​m/Screenshot_2015-07-19_06-06-32.png
20:18 VanessaE (better example)
20:18 kilbith yes, maybe same problem
20:18 VanessaE I wonder if it's related to bug that causes thousands of small "holes" in the mesh?
20:19 RealBadAngel i can make it smaller, less visible by making mesh having more cubes
20:19 RealBadAngel but at the price of speed
20:19 selat joined #minetest-dev
20:20 RealBadAngel wieldmesh.cpp line 37
20:20 RealBadAngel #define MIN_EXTRUSION_MESH_RESOLUTION 16
20:21 RealBadAngel change here 16 to lets say 256
20:21 RealBadAngel and recompile
20:21 RealBadAngel your dots shall be gone
20:22 RealBadAngel btw, i dont have such dots here. somehow its depending on video driver too
20:22 RealBadAngel propably depending on floating point accuracy when rendering
20:22 kilbith that's possible too
20:25 RealBadAngel i wonder if setting 512x just for hand, upscaling then a texture to fit the mesh would do the trick
20:25 RealBadAngel ofc with upscaled texture i could apply the filters
20:25 VanessaE bit of a brute-force way to go about it, isn't it?
20:26 kilbith well that's just cosmetic anywats
20:27 RealBadAngel cosmetic or not, but a bug
20:27 TBC_x It's not that annoying
20:27 TBC_x at night
20:27 RealBadAngel yet we dont have a working replacement for it
20:27 zat joined #minetest-dev
20:28 TBC_x make night permanent
20:28 TBC_x problem solved :)
20:28 RealBadAngel lol
20:28 TBC_x next
20:28 RealBadAngel i know the better solution
20:28 RealBadAngel geometry shaders
20:28 kilbith that's more important bug : https://github.com/minetest/minetest/issues/2350
20:28 kilbith (for the wieldhand that is)
20:28 RealBadAngel emit vertex and copy the pixels
20:29 RealBadAngel but thats for shaders only
20:30 VanessaE you know, when I look at that hand and bucket, something comes to mind:  I know minetest (or irrlicht?) can't properly deal with ngons that are concave - it tries to "fill-in" the space with additional triangles (or just re-maps what it has)./
20:30 VanessaE those images look like exactly the same effect
20:30 VanessaE kahrl: ^^^
20:32 kilbith VanessaE: on Windows this happens on everything that's wielded
20:32 kilbith on linux it's ok
20:33 kilbith so the determining factor may be the drivers used
20:35 RealBadAngel i havent said that. hmmm would kill me for that ;)
20:35 RealBadAngel lol
20:41 TBC_x ok, #2946 is waiting for jenkins and friends
20:41 ShadowBot https://github.com/minetest/minetest/issues/2946 -- Fixed minimap memory leak by t0suj4
20:42 paramat left #minetest-dev
20:49 Robert_Zenz joined #minetest-dev
20:52 TBC_x anyone have any objections to the pr?
20:53 leat1 joined #minetest-dev
20:55 blaise joined #minetest-dev
20:58 RealBadAngel TBC_x, in fact you can push it to cover the leaks. what im coding for minimap is gonna take me a week or more even
20:58 TBC_x it's sitting there
20:58 TBC_x not gonna do any modifications
20:58 TBC_x recompiling irrlicht & mt
20:59 RealBadAngel i will move minimap outta mapblock mesh anyway
20:59 ElectronLibre joined #minetest-dev
20:59 RealBadAngel updates will be triggered the same place, but threads will be separated
21:01 TBC_x I've noticed that server is sending mapblocks before the mapgen lands finishing touches
21:08 TBC_x It is unbelievable how swift is my compiling without address/undefined sanitizers
21:09 TBC_x s/my/mt/
21:16 hmmmm so does this mean the minimap is going to finally get cleaned up?
21:20 TBC_x I still can't find out how the refcouting is failing
21:23 TBC_x wow, such harsh words
21:27 RealBadAngel hmmm, ^^
21:27 kilbith you missed one 'm'
21:28 RealBadAngel to get something better you have to get it in first and got it tested by more users
21:28 RealBadAngel thats is the meaning of "development branch"
21:30 RealBadAngel also nobody has ever wrote a piece of code that was perfect from the very begining
21:33 hmmmm thank goodness
21:34 hmmmm RealBadAngel, I am glad that you're working on improving it
21:37 RealBadAngel hmmmm, also idk if you have read all the logs. i have to admit you were right that mapblock_mesh is not a proper place for minimap calls. its called way to often
21:38 hmmmm not only that, but did you profile the amount of time the scanning takes?
21:38 hmmmm meshmaking takes about ~500us, the scan takes ~130us
21:38 RealBadAngel its not the complete scan, its just cache data feed
21:38 hmmmm that's a 25% increase
21:39 hmmmm I know, the initial scan
21:39 Routh joined #minetest-dev
21:39 hmmmm this means
21:39 RealBadAngel main place i WANT improved code is getPixel
21:39 RealBadAngel translations between block coords and node's takes way too much
21:40 RealBadAngel its being made 256x256 times
21:40 hmmmm if the original scan took 450-500 us and the new one takes 600-650 us, that's a 25% performance hit on meshmaking throughput
21:40 hmmmm ideally there'd be only one scan
21:41 hmmmm no scan once, then cache in a 3d position, then scan again
21:41 RealBadAngel i tried to limit it
21:41 hmmmm why does it need to be cached in a 3d space if it's only 2d?
21:41 RealBadAngel but results are deleted in a queue
21:41 hmmmm (i realize you do this for caves and stuff, but there is a more clever algorithm)
21:41 RealBadAngel look, theres main problem with mesh making
21:41 hmmmm it's called too often
21:41 hmmmm yes I get that
21:41 RealBadAngel when block arrives it triggering an update
21:42 RealBadAngel but not only for itself
21:42 RealBadAngel but with edges
21:42 RealBadAngel each one of them is causing 7 updates
21:42 RealBadAngel thats a shitload
21:42 TBC_x is that because of lightning?
21:43 hmmmm literally anything
21:43 hmmmm the cursor moving to a new block
21:43 hmmmm a crack being formed in the current block
21:43 RealBadAngel textures connecting, backface culling, lighting, liquids spread
21:43 hmmmm a change in lighting
21:43 hmmmm instead of talking about it we could be fixing it
21:43 hmmmm i'm not going to babble on about this.  later
21:44 TBC_x is lock order inversion a good thing?
21:44 RealBadAngel hmmm to get it fixed we have to move out of it literally everything
21:44 RealBadAngel cracks, animations, highlighting
21:44 hmmmm right
21:44 RealBadAngel only pure mesh shall stay here
21:44 hmmmm TBC_x:  does that happen somewhere?
21:44 hmmmm it's not a good thing, but it's not necessarily bad
21:45 Calinou joined #minetest-dev
21:45 TBC_x somewhere in createOpenALSoundManager
21:46 TBC_x hmm
21:46 TBC_x it happens in glib
21:46 TBC_x If I am interpreting the results correctly
21:47 hmmmm we don't use glib and i don't think OpenAL does either
21:47 hmmmm i don't think we have any dependency on glib for that matter
21:50 TBC_x #1 g_static_rec_mutex_lock  #2 createOpenALSoundManager
21:51 TBC_x #3 GUIEngine::GUIEngine
21:52 TBC_x so far it looks like false alarm
21:52 hmmmm i dunno man
21:52 hmmmm http://www.freshports.org/audio/openal-soft
21:52 hmmmm the only dependency it seems to have is devel/cmake
21:53 TBC_x have to dig in
21:54 hmmmm % pkg info -d openal-soft-1.16.0_2
21:54 hmmmm openal-soft-1.16.0_2:
21:54 hmmmm could it be that your openal was compiled with pulseaudio support
21:55 TBC_x yeah
21:55 TBC_x that's probably the case
21:55 TBC_x let's ignore that
21:57 RealBadAngel speakin of which, there was an issue somwhere? pr? positional audio
21:58 RealBadAngel mumble
21:59 Calinou yes, there's an issue on Mumble positional audio support
21:59 RealBadAngel it seems damn easy to add
21:59 kilbith i remember you wanted to integrate mumble in-game, RBA
22:06 RealBadAngel https://www.youtube.com/watch?v=CHMjBPUmGZk
22:06 RealBadAngel theres a vid on mc using mumble
22:07 RealBadAngel whats nice in mumble it needs only a pos
22:07 RealBadAngel what reminded me of mumble: request on get player velocity
22:08 RealBadAngel if player is falling down he can start to scream
22:08 RealBadAngel if you are using mumble you can hear scream incoming
22:08 RealBadAngel just before the splash ;)
22:12 AnotherBrick joined #minetest-dev
22:14 sloantothebone joined #minetest-dev
22:14 TBC_x can I enable sw rendering?
22:15 TBC_x for irrlicht
22:15 TBC_x I mean, is it supported?
22:16 RealBadAngel dont even try it
22:16 RealBadAngel its a joke
22:16 paramat joined #minetest-dev
22:17 TBC_x well... better than GL_OUT_OF_MEMORY
22:18 paramat sfan5 pinetree updated, i'll push game#587 soon
22:18 ShadowBot https://github.com/minetes​t/minetest_game/issues/587 -- Default: Add improved mgv5/mgv7 tree schematics by paramat
22:20 RealBadAngel TBC_x, opengl is a standard. we shall think on makin it default and only way to display things
22:21 RealBadAngel 2.1 is nine years old
22:21 RealBadAngel and its enough to make sure most basic shaders will work
22:21 RealBadAngel ie lighting stuff, fog and even bumpmapping
22:22 RealBadAngel we shall drop everything else imho
22:22 kilbith sfan5 wants that too
22:23 RealBadAngel nowadays if a device is not able to use opengl it usually means its a microvave or a washin machine
22:23 RealBadAngel but thats not for sure, some models can have support for it ;)
22:23 proller joined #minetest-dev
22:24 RealBadAngel ive already saw a watch running minetest
22:24 TBC_x rofl
22:24 kilbith https://www.youtube.com/watch?v=6MItj9AJbP4
22:26 Calinou +1 for dropping everything non-OpenGL/GL ES
22:27 Calinou software rendering is plain obsolete, and Direct3D is broken
22:27 RealBadAngel having shaders (basic version) a default have some more pluses
22:28 RealBadAngel final color blend for whole the scene
22:28 TBC_x that's ridicilous
22:28 RealBadAngel keeping animated textures in sync
22:28 RealBadAngel TBC_x, its not
22:28 Calinou we could have mandatory shaders
22:28 RealBadAngel shaders are invented more than 10 yrs ago
22:28 Calinou that means we can deprecate the fixed-function pipeline
22:28 TBC_x was talking about the watch
22:29 RealBadAngel lol
22:29 Calinou Sauerbraten and Red Eclipse dropped fixed-function recently
22:29 Calinou however we must make sure the shaders run on Android
22:29 RealBadAngel Calinou, i can do that
22:29 TBC_x and on watch
22:29 TBC_x :)
22:29 RealBadAngel if i do keep it on 2.1 level it can work
22:29 Calinou however the legacy crowd will whine :(
22:29 TBC_x they get software rendering, somehow
22:29 Calinou I can't think of anyone whose device doesn't support shaders here
22:30 RealBadAngel even that intel 945g is able to run 2.1
22:30 RealBadAngel so we are secure
22:30 Calinou requiring 2.1 sounds OK, it's what Sauerbraten/Red Eclipse do
22:30 Calinou even Minecraft's doing it now
22:30 RealBadAngel i wanted to split shaders into three groups
22:31 RealBadAngel basic -- only lightighin, fog and waving
22:31 Calinou we could have a shader-based fog
22:31 RealBadAngel lite - above + bumpmapping and autogen
22:31 Calinou so that it wraps nicely around the player, and can be volumetric
22:31 RealBadAngel full - relief mapping and whatever we can imagine
22:32 kilbith s/lite/normal
22:32 TBC_x So how do I get software rendering going?
22:32 RealBadAngel you really want to see that  horror?
22:33 TBC_x I don't care as long as I don't get segfaults
22:33 RealBadAngel im afraid if you will be able to sleep again after experiencin it
22:34 RealBadAngel i cant take such responsibility
22:34 TBC_x tsan is quite picky
22:34 TBC_x and it won't allow opengl
22:36 RealBadAngel we dont need anything other than GL/GL ES
22:37 RealBadAngel DX is platform specific and thus unreliable
22:37 RealBadAngel theres simply no other option
22:37 RealBadAngel software rendering is also a mistake
22:38 RealBadAngel see how lighting works when being care by cpu not the shaders per pixel
22:39 TBC_x I don't care, I want sw rendering
22:42 RealBadAngel excuse me asking, are you a masochist? ;)
22:43 TBC_x whatever
22:43 TBC_x just tell me how
22:43 TBC_x I won't be looking at it
22:43 TBC_x I promise
22:44 RealBadAngel just visit settings tab and choose a renderer
22:44 RealBadAngel then restart mt
22:45 kilbith TBC_x: rtfm : https://github.com/minetest/minetest/blo​b/master/minetest.conf.example#L127-L129
22:45 RealBadAngel i just did that. omfg
22:45 RealBadAngel bloody hell
22:45 RealBadAngel TBC_x, apage satanas :P
22:47 TBC_x Damn... expected It would be better hidden away
22:47 TBC_x not in a settings menu
22:47 RealBadAngel rotfl
22:47 RealBadAngel i was against putting it there
22:48 RealBadAngel its plain dumb
22:48 kilbith so RBA, 1) take off that dropdown 2) add a dropdown for shaders
22:49 kilbith lot of place gained
22:49 RealBadAngel kilbith, an external app based on mt and irr is a good idea
22:50 RealBadAngel you dont have to run it too often
22:50 kilbith that's a long term goal
22:50 RealBadAngel propably once or twice to get your game fine tuned
22:50 RealBadAngel after that all you need is an nice screen and go button
22:51 RealBadAngel with MUSIC ofc and for christ sake too ;)
22:52 RealBadAngel since im former demoscene coder i think i will write one day a thing that will replace that ugly shit we can all see launching the game
22:54 RealBadAngel one of the greatest composers for demoscene granted us the rights to use his music for our project too
22:55 RealBadAngel we are just constantly shooting our feets imho
22:56 RealBadAngel https://www.youtube.com/watch?v=cjHNPzv9sGY
22:57 kilbith 173 KB for 4.25, that's pretty well compressed
22:57 kilbith *4:25
22:57 RealBadAngel soundtracker
22:58 RealBadAngel i wrote an implementation of it for z80 and A/D made out of bunch of resistors ;)
22:59 RealBadAngel i mean D/A ofc
23:00 TBC_x oh
23:01 kilbith talking about MT on ridiculous devices..
23:01 kilbith even better : https://www.youtube.com/watch?v=4WKOFg0vCwc
23:02 RealBadAngel lol i saw i already
23:02 RealBadAngel flawless
23:02 RealBadAngel thats what i admire the most
23:02 paramat an external app is bad because it is external, MT launch is not 'ugly shit', you're getting carried away with your ideas again and trying to put down MT
23:03 RealBadAngel paramat, what we do have is neither simple, nice or easy to maintain
23:03 RealBadAngel theres no place for basic settings
23:03 paramat sure it can be improved
23:04 RealBadAngel and looks like done 15 yrs ago
23:04 RealBadAngel wake up
23:04 RealBadAngel its 21st century
23:04 paramat retro/8 bit is cool, and MTs character
23:04 RealBadAngel youre making a video GAME
23:04 kilbith it's not terasology either
23:04 RealBadAngel ofc not
23:05 RealBadAngel terasology is quite nice but went wrong direction
23:05 kilbith it's a fuckin gas industry
23:05 RealBadAngel java? blurp
23:05 TBC_x Just kick the user right into the 'menu' world
23:06 RealBadAngel i will just shut up and code nice, up to the current century count menu
23:06 RealBadAngel before finishing i wont show you anythin
23:07 werwerwer joined #minetest-dev
23:08 RealBadAngel i do know still artists from the demoscene
23:08 RealBadAngel most of them are still in the business
23:09 RealBadAngel i was suprised lately that guy that wrote music scores for my 8 bit game, wrote a music for polish president reelection campaign
23:16 RealBadAngel anyway, i will just do it
23:26 TBC_x whoops
23:26 TBC_x I looked at it
23:48 TBC_x client http://sprunge.us/cdFP
23:48 TBC_x server http://sprunge.us/FWIB
23:49 TBC_x also, there are some false positives
23:50 TBC_x mostly when it is somewhere around  #0 JThread::TheThread
23:50 TBC_x due to a spinlock
23:53 est31 joined #minetest-dev
23:56 paramat now pushing game 587
23:58 est31 RealBadAngel, the problem with menu music is that minetest is a sanbox game. We have to find a generic enough piece which isnt shit :)
23:59 est31 but the mercury rain piece is better imo than that medieval styled piece from your pr
23:59 est31 only personal opinion
23:59 est31 also, perhaps we should have "extended settings" menu, reachable by button from settings tab.

| Channels | #minetest-dev index | Today | | Google Search | Plaintext