Minetest logo

IRC log for #minetest-dev, 2014-04-14

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

All times shown according to UTC.

Time Nick Message
00:10 salamanderrake joined #minetest-dev
00:15 ShadowNinja Here's a more complete formspec mockup: http://ix.io/bEZ/lua
00:16 ShadowNinja If we use full Lua mixed array/map serialization we can make t[1] a synonym for t["type"] and make it simpler.
00:17 ShadowNinja Or smaller rather.
00:21 iqualfragile_ joined #minetest-dev
00:27 ShadowNinja I notice that the ASync API functions are passing str:len() after str to C a lot.  Seems that sapier didn't know about the lua_getstring functions with a size.
00:53 kahrl joined #minetest-dev
01:11 Taoki joined #minetest-dev
01:46 luizrpgluiz1 joined #minetest-dev
01:46 luizrpgluiz1 left #minetest-dev
04:21 cheapie joined #minetest-dev
04:52 ShadowNinja ~tell sapier I removed the marshal dependency from the async API using string.dump and loadstring.  I also fixed code style and a few real bugs (eg, missing lua_pop)  Is the broken LuaJIT beta check still relevant?  Experimental patch: http://ix.io/bF7/diff
04:52 ShadowBot ShadowNinja: O.K.
05:09 ShadowNinja Er, had other changes there.  http://ix.io/bF8
05:12 ShadowNinja I also found a bug.  Pausing the game before the world is visible will prevent it from loading.
05:40 SmugLeaf joined #minetest-dev
05:42 darkrose joined #minetest-dev
05:42 darkrose joined #minetest-dev
05:43 werwerwer joined #minetest-dev
07:19 ShadowNinja And yet again my C Lua changes have exploded into a monster: 24 files changed, 721 insertions(+), 1515 deletions(-)
07:20 ShadowNinja I added the errfunc at init, rather than as every C -> Lua call.
07:20 ShadowNinja at*
07:21 ShadowNinja I also mv'd lua_api/l_async_events.* to cpp_api/s_async.*  Since it actually belongs there.
07:23 ShadowNinja http://ix.io/bFd/diff
07:24 ShadowNinja And with that I'm off.  o/
07:59 PenguinDad joined #minetest-dev
09:01 darkrose joined #minetest-dev
09:01 darkrose joined #minetest-dev
09:43 seolfor joined #minetest-dev
09:44 seolfor can somebody help me with cmake config ? i can't add FrreType libarys
10:08 proller joined #minetest-dev
10:13 restcoser joined #minetest-dev
10:13 troller joined #minetest-dev
10:17 proller joined #minetest-dev
10:26 proller joined #minetest-dev
10:33 proller joined #minetest-dev
11:29 rdococ joined #minetest-dev
11:37 ImQ009 joined #minetest-dev
11:55 kaeza joined #minetest-dev
12:04 Jordach joined #minetest-dev
12:42 hmmmm joined #minetest-dev
13:03 Shardvex joined #minetest-dev
13:12 PilzAdam joined #minetest-dev
13:49 Shardvex joined #minetest-dev
14:04 PenguinDad joined #minetest-dev
14:13 Shardvex joined #minetest-dev
14:23 Garmine joined #minetest-dev
14:28 smoke_fumus joined #minetest-dev
14:36 twoelk joined #minetest-dev
14:53 proller joined #minetest-dev
14:53 BlockMen joined #minetest-dev
15:07 seolfor joined #minetest-dev
15:11 Taoki[laptop] joined #minetest-dev
15:12 seolfor BlockMen, darkrose, PilzAdam will somebody help me with cmake config? I can't add FreeType libaryr into cmake config
15:12 seolfor libarys*
15:15 Shardvex #minetest seems to be more active right now
15:16 seolfor thx for information Shardvex, but mayby you known why i cant add libary path of freeType into cmake config ?
15:27 BlockMen seolfor, could you be more concrete what is not working?
15:31 zat joined #minetest-dev
15:32 RealBadAngel joined #minetest-dev
15:32 RealBadAngel hi
15:32 seolfor i made minimal build with vs2012 configued with cmake
15:33 PilzAdam RealBadAngel, fix your shaders: https://github.com/minetest/minetest/issues/1193
15:33 seolfor after this i tried to add freetext dependency
15:33 RealBadAngel PilzAdam, im aware of it
15:34 PilzAdam also water is opaque sometimes and light has no effect on it
15:34 rubenwardy joined #minetest-dev
15:34 RealBadAngel the question is i shall push changes that those bugs do not apply or 1st change older code
15:34 PilzAdam what?
15:34 RealBadAngel and then push newer code
15:35 RealBadAngel i have rebuilt shaders completely meanwhile
15:35 PilzAdam fix the bugs without introducing new bugs
15:36 PilzAdam also why do you keep rewriting shaders completly?
15:36 RealBadAngel not the shaders really, way of handling them
15:37 RealBadAngel i have moved generation of shaders from realtime to startup
15:38 RealBadAngel simple step, but i do have 40-60 fps again, with all the effects turned on
15:39 RealBadAngel and i mean with normals generation on the fly
15:41 RealBadAngel also, this rewrite means one shader for nodes
15:41 RealBadAngel and others for postprocess
15:43 RealBadAngel again, the question, shall i fix the old code or push the new one....
15:44 RealBadAngel because the bug is kinda silly, and openGL related
15:47 RealBadAngel on one machine code vec4 (color.rgb, alpha) is proper on others vec4 (color.r, color.b, color.g, alpha) is needed
15:48 RealBadAngel this is kinda sick
15:49 RealBadAngel that one above caused light issues for water at night
15:49 seolfor RealBadAngel, can't you be more specific with on is on with machine ? (vec4 ) mayby ther is some macro needed
15:50 Calinou joined #minetest-dev
15:51 seolfor im not shure what "diffrent" machines you mean ... linux/windos x32/x64, or any random thing ..
15:53 ShadowNinja seolfor: Different versions of OpenGL or graphics cards.
15:53 RealBadAngel genau
15:53 RealBadAngel same code behaves very different way
15:54 RealBadAngel or even wont compile
15:54 RealBadAngel gl is a bitch
15:54 RealBadAngel and they call it a standard, LOL
15:56 Calinou joined #minetest-dev
15:56 seolfor if thats the case so it's realy a quite problem
15:59 seolfor if it would be only diffrent version of openGL it whould be not so bad, but if grahpic card's make inpact on it's I whould consider it as bug of OpenGL
15:59 NakedFury joined #minetest-dev
16:05 RealBadAngel i have to be very carefull what im using, have to check what works for nvidia what for ati
16:05 RealBadAngel this is ridiculous
16:05 seolfor this is Open software ;]
16:06 RealBadAngel actually not
16:06 seolfor realy ? OpenGl is not public code ?
16:06 RealBadAngel thats why Thorvalds show nvidia a finger
16:07 RealBadAngel interface to hardware is closed
16:08 seolfor so the "best" part is closed .. what a lame :p
16:08 RealBadAngel you can just quess whats right
16:09 RealBadAngel open source drivers are reverse engering
16:09 Calinou <RealBadAngel> interface to hardware is closed
16:09 Calinou like in any brand
16:10 Calinou <RealBadAngel> thats why Thorvalds show nvidia a finger
16:10 RealBadAngel WE shall not let it happen
16:10 seolfor ;D
16:10 Calinou he retracted his finger a few months ago when NVIDIA started pushing straw man patches
16:11 RealBadAngel we do have much more fingers aviable
16:12 RealBadAngel because if we will let it pass, we soon will be coding in pauses for watchin the commercials
16:14 sfan5 ShadowNinja: can I have your threadname patch again? (the non-hacky one)
16:16 ShadowNinja sfan5: The non-hacky one doesn't work with Windows at all.  The hacky one works only for MSVC.  And the hacky part is only the MSVC section.
16:16 sfan5 how do you know it does not work at all?
16:16 ShadowNinja You can get the patch from my PR.  Because SetThreadName doesn't exist on any version of Windows.
16:21 sfan5 BTW #1224 should be merged
16:21 ShadowBot https://github.com/minetest/minetest/issues/1224 -- Fix problem with newer MinGW runtimes by sfan5
16:22 sfan5 it fixes compiling on my freshly compiled MinGW gcc 4.8
16:22 seolfor guys why i can makde Debug build but i get linking errors in Release one ?
16:22 seolfor make*
16:23 seolfor and please dont tell me somthing like "instal linx"
16:33 seolfor im using newest version of minetest, i made a debug build, run it and in attempt of world creation i getting ingame Error:Invalid game index - it's only for me or newest commit is broken ?
16:33 sfan5 seolfor: dl yl
16:33 PilzAdam RealBadAngel, opengl is an API specification; it has nothing to do with code
16:34 sfan5 seolfor: please paste the errors you get
16:35 seolfor sfan5, like i sad i got "Error: Invalid game index"
16:35 sfan5 I mean the linking errors
16:37 seolfor http://pastebin.com/e4MqnVDH,
16:37 jin_xi joined #minetest-dev
16:42 seolfor sfan5, any idea ?
16:42 sfan5 hm
16:43 sfan5 no idea
16:43 seolfor ;] thx any way, and how about this one ingame error ?
16:43 ShadowNinja Seems MSVC and MT both define _str*.
16:45 seolfor im seek and tired of this msvc.. i will work only in debug mode for now
16:47 seolfor any ide'a what is this ?
16:47 seolfor http://snag.gy/KntV3.jpg
16:49 sfan5 your irrlicht was compiled without support for other fonts
16:50 seolfor sfan5, so i have to do what? compile irrlicht or add gettext font lib into project ?
16:50 sfan5 gettext is for translations, not fonts
16:51 sfan5 either compile irrlicht or use sfan5.pf-control.de/irrlicht-1.8.1-win32.zip
16:51 seolfor thx for pack
16:51 sfan5 I have some more libs on thre
16:51 sfan5 there*
16:52 seolfor it's compiled with gcc ? i'm using vs2012 it will work ?
16:53 sfan5 these: http://sfan5.pf-control.de/l​uajit-2.0.1-static-win32.zip & http://sfan5.pf-control.de​/libleveldb-1.12-win32.zip & http://sfan5.pf-control.de/freetype-2.3.5-1.zip & http://sfan5.pf-control.de/gettext-0.14.4.zip
16:53 sfan5 it was compiled with gcc, but it should work with msvc
16:53 Shardvex joined #minetest-dev
16:53 seolfor thx
16:53 seolfor i'm missing only gettext rest i made alredy
16:53 rubenwardy left #minetest-dev
16:55 seolfor and i had already freetype in my project it's it font pack ? or am i missing dll in build folder ?
16:55 sfan5 if you have a .dll missing it would not start
16:57 seolfor yep i added them already but im curious why i dont got fonts .. (i have to install some dictionary checking into irc ;/  )
17:01 seolfor sfan5, if you want i have also curl libogg/vorbis copiled libs
17:02 seolfor it may be usefull for somebody
17:02 sfan5 I have all libs in compiled form
17:02 sfan5 +alreadyy
17:02 sfan5 -y
17:02 sfan5 I provide windows builds
17:02 seolfor ah is that so .. ;]
17:05 rubenwardy joined #minetest-dev
17:07 Ritchie_ joined #minetest-dev
17:09 Megaf joined #minetest-dev
17:09 seolfor sfan5, i can't link your irrlicht libary ;(
17:10 seolfor so it's looks like i have to compile it myslef
17:10 Megaf Hi everyone
17:17 seolfor sfan5, this compilation of irrlicht mast be done with some speclal parameters or linking's for fonts supprot ?
17:17 sfan5 no
17:19 seolfor just compile it right? ok i will do it (but first it's looks like im missing directX sdk)
17:31 sfan5 you do not need the DX sdk if you don't need DirectX support
18:01 Shardvex joined #minetest-dev
18:02 seolfor sfan5, i need sdk dx for compilation process
18:03 seolfor +irrlich
18:09 sfan5 merging #1224 in 10 mins
18:09 ShadowBot https://github.com/minetest/minetest/issues/1224 -- Fix problem with newer MinGW runtimes by sfan5
18:27 RealBadAngel PilzAdam: it seems that you have never code anything open GL
18:28 RealBadAngel some code is valid for ati, some for nvidia, not to mention revisions...
18:29 RealBadAngel opengl is a joke
18:30 RealBadAngel most fucked up programming languange i ever saw
18:30 sfan5 opengl is not a programming language
18:31 RealBadAngel you know what im talkin about
18:31 sfan5 the shader lang?
18:31 RealBadAngel up
18:31 RealBadAngel yup
18:32 PilzAdam RealBadAngel, I think _you_ dont know what you are talking about :-p
18:32 seolfor PilzAdam, im shure he mean API instead "language"
18:32 RealBadAngel PilzAdam: for christ sake, code something once
18:33 RealBadAngel instead of fooling around
18:33 RealBadAngel any other case, just shut the up
18:33 seolfor ;] i wana code somthing .... instead of this fight with compitalione on win
18:34 Exio4 joined #minetest-dev
18:34 RealBadAngel *fuck
18:34 sfan5 seolfor: compile with mingw if you don't want to mess with stuff
18:35 seolfor sfan5, can u send me ALL your dependencies and compilation projects etc ?
18:35 RealBadAngel anyway, fixes for the issues will come this weekend
18:36 seolfor to hell with it >.> im instaling linux
18:36 RealBadAngel propably together with new architechture
18:37 RealBadAngel going back is not possible
18:37 sfan5 seolfor: compiling works pretty good on windows
18:37 sfan5 but only when you use mingw
18:37 seolfor sfan5, i tried
18:37 sfan5 with MinGW?
18:37 seolfor i made all dependencies with mingw
18:37 sfan5 compile minetest with mingw too
18:38 seolfor but i wasnt be able to compile minetest
18:38 sfan5 post an error log then
18:38 RealBadAngel im encountering much more problems fixing stuff than doing them properly from scratch
18:39 seolfor hmy .. then.. can you give me all depenedencies for mingw again ? i removed my compilations already ..
18:39 sfan5 ok
18:39 seolfor for mingw ofcourse*
18:43 seolfor sfan5, are you there ? i realy started thinking about linux
18:43 sfan5 yes
18:43 sfan5 uploading..
18:44 seolfor oh, sory i will w8 then ;D
18:45 sfan5 seolfor: sfan5.pf-control.de/packages-minetest-win32.7z
18:46 seolfor asome.. thx a lot i will give them a shot ;]
18:50 seolfor sfan5, about OGG_LIBRARY with one should i include ? /libogg-1.2.1-dev/lib/libogg.a or *.dll.a ?
18:51 sfan5 if you don't want to put the .dll in your bin folder use the .a otherwise the .dll.a
18:51 sfan5 this is a more or less general rule
18:51 seolfor thx
18:54 tomreyn joined #minetest-dev
18:59 seolfor sfan5, i already made "generation" form cmake for mingw, and now how do i compile minetest ?
19:00 sfan5 run 'make'
19:00 seolfor from msys console ?
19:00 sfan5 yep
19:00 seolfor i done it but ..
19:01 sfan5 paste the error
19:02 seolfor yep w8, .... bwt should i run it form main minetest folder or form cmake generated ?
19:03 seolfor sfan5, http://snag.gy/aRotn.jpg
19:03 sfan5 from the folder cmake made
19:04 seolfor so im done it ..
19:04 sfan5 where is your 'Makefile'?
19:04 seolfor but nothing heppend (or it's seems like msys closed after i typed make)
19:05 seolfor in .. .. . minetest\bin folder
19:05 seolfor path is right
19:06 sfan5 hm
19:06 seolfor i hate "cmd" in windows >.< right now i have 4 diffrend comand shells >.<
19:06 sfan5 run cmake with the build dir as a seperate directory, e.g.  minetest\build
19:07 seolfor i run cmake for destination directory minetest\bin - it shouldnt be much diffrence
19:07 seolfor mayby my mingw is coruppted ?
19:08 sfan5 I don't think so
19:08 sfan5 wait wait
19:08 seolfor but it's looks aloust like shell clrash after typing cmake
19:08 sfan5 <sfan5> run cmake with the build dir as a seperate directory, e.g.  minetest\build
19:08 sfan5 try that
19:08 seolfor ok
19:09 seolfor o0
19:09 seolfor i think i found problem ...
19:11 seolfor last one i made for mingw makefile http://snag.gy/xzQtD.jpg
19:11 seolfor instead of msys makefiles
19:13 seolfor sfan5, but now i get http://0bin.net/paste/CX8iZRrvcgh4KWjZ#, so cmake should be done for mingw or msys ?
19:13 sfan5 ...
19:13 sfan5 umm
19:13 sfan5 the paste link does not work
19:14 seolfor http://ctrlv.it/id/Mzc3OTc1
19:15 sfan5 hm
19:15 sfan5 try mignw
19:15 sfan5 mingw*
19:15 sfan5 maybe you need to download make seperatly and run it from normal cmd
19:16 seolfor sfan5, done http://ctrlv.it/id/Mzc3OTc2
19:16 sfan5 that looks good
19:19 seolfor ah .. it's comping now .. i run make form cmd shell like : C:\MinGW\bin\mingw-make, and it failed already ;( but how i got at least some error
19:20 seolfor error: ft2build.h: No such file or directory
19:20 seolfor freetype woring linking ?
19:25 seolfor yep freetype variables are not showing in cmake-gui for shure
19:26 sfan5 freetype has 2 different includes
19:26 sfan5 http://dev.minetest.net/CMake_Options#Freetype2
19:26 seolfor i added them already
19:27 sfan5 ^ seolfor: consult that wiki page for information
19:28 seolfor but realy why ther are two includes ;] *_ft2build *_freetype2 it's not normall at all ..
19:28 seolfor how i have to figure out how to debug this build ;)
19:29 celeron55 seolfor: i have really bad experiences trying to use cmake with msys
19:29 seolfor sfan5, mingw by defauld is making debug or release build ?
19:29 celeron55 it just doesn't understand what it's running on
19:29 celeron55 run cmake on the regular windows command prompt and tell it to use mingw and it will work
19:30 sfan5 seolfor: cmake makes release by default
19:30 celeron55 and remember to clear all of cmake's cache (CMakeCache.txt, CMakeFiles)
19:30 sfan5 seolfor: there were always two includes
19:30 seolfor celeron55, i done it already exacly like you say
19:35 seolfor error at compiling proces .. http://ctrlv.it/id/Mzc3OTc3
19:35 seolfor at 100% ;]
19:40 sfan5 hm
19:40 sfan5 disable leveldb for now
19:40 sfan5 seems like I did something wrong
19:41 seolfor i did it and now recoplinig...
19:47 seolfor sfan5, again failed http://ctrlv.it/id/Mzc3OTg4 and now i don't have any idea why
19:48 sfan5 o.O
19:48 sfan5 which mingw version do you have?
19:49 seolfor hmy .. mingw-make is 3.82.90
19:53 seolfor sfan5, any idea ?
19:53 sfan5 which mingw-gcc version?
19:54 seolfor 4.8.1
19:58 ShadowNinja celeron55: Did you see my SCons minetestmapper patch?
19:59 ShadowNinja I made another patch for minetestmapper that tripples speed: http://pastebin.ubuntu.com/7251559/
19:59 seolfor sfan5, can you give me your cmake config ? i will only change dependencies paths
19:59 sfan5 eh, sure
20:00 seolfor becouse this is simple seek .. >.<
20:00 ShadowNinja Also, comments on http://ix.io/bFd/diff ?
20:01 seolfor o .. i have better idea i will not chage anything in cmake config i will simply move dependencis
20:02 sfan5 seolfor: http://0bin.net/paste/YrCDZng3IjNpfGCG#p9h​o/WHZ1cn9+V9axeIu6LWFax6h4ZqGS/64L0XMIJA=
20:03 seolfor "simply" <- have new meening for me now >.> like "i will not use door to get in, i will make another one exacly the same like sfan5 got, and then we will se if they will work" ;(
20:07 ShadowNinja Also, who came up with this pythonmodulo bit twidling?  http://ix.io/bG2/diff  I'll push that in a minute.
20:09 seolfor sfan5, what is it : /home/stefan/mt-build/buil​dbot/toolchain_mingw.cmake
20:09 seolfor and this : /usr/bin/msgfmt
20:09 sfan5 msgfmt is used for gettext
20:09 sfan5 just ignore the toolchain_mingw.cmake file
20:10 seolfor and this msgfmt ?
20:13 sfan5 just disable gettext for now
20:16 ShadowNinja The Minetest client and server-only build use seperate project names, meaning that everything that the server uses is build twice.  Can this be done in a better way?
20:23 ShadowNinja sfan5: Comments on my mapper patch?
20:31 cheapie joined #minetest-dev
20:35 BlockMen left #minetest-dev
20:36 proller joined #minetest-dev
20:39 seolfor sfan5, to hell with windows, still not working tomorow i will my lovly archlinux
20:40 seolfor i will install my lovly*
20:43 tiemay joined #minetest-dev
20:45 ShadowNinja Grrr, that noise bug is making Minetest unusable.  hmmmm: <-- Has nobody else experienced this?
20:49 grrk-bzzt joined #minetest-dev
20:50 VanessaE_ what noise bug?
20:53 ShadowNinja VanessaE_: resizeNoiseBuf throws an std::bad_alloc.  Release builds too (although a catch (...) catches it and prints it rather than crashing)
20:53 ShadowNinja Oddly only my development installation has it.
20:53 ShadowNinja Possibly a mod.
20:54 VanessaE_ hm.  I haven't been playing much lately so I haven't seen anything unusual like that.
20:55 tiemay I just created a minetest cleaner for bleachbit ( https://github.com/az0/cleanerml/​blob/master/pending/minetest.xml ), and it needs to be reviewed before getting added to the release. You all probably know minetest best, so it would be nice if you helped ( http://bleachbit.sourcefor​ge.net/contribute/cleaner )
20:55 ShadowNinja Happens even with all mods disabled.
20:56 ShadowNinja VanessaE_: I've had this issue for a while.
21:11 grrk-bzzt joined #minetest-dev
21:23 hmmmm ShadowNinja, huh?
21:24 hmmmm the noise has been in there and hasn't changed for about a year and a half now
21:24 hmmmm nobody else has ever had this problem.....
21:24 celeron55 tiemay: what the mod_settings one isn't officially supported by minetest; the behavior is completely undefined if you do that
21:24 hmmmm std::bad_alloc typically means you're trying to allocate way too much, so chances are a mod has a value that's really messed up
21:25 celeron55 tiemay: probably breaks the world
21:25 hmmmm I don't do error checking on the noise parameters, sorry.  I really should though I admit
21:25 celeron55 s/what the mod_settings one/what the mod_settings one does/
21:25 hmmmm my code has a certain philosophy though, and users of Noise are able to change most of the parameters after the noise has already been constructed, without "setter" method overhead
21:26 hmmmm the downside of this philosophy is that it has a certain amount of trust
21:26 celeron55 tiemay: the four others are fine
21:26 tiemay celeron55: I tested it, currently the file just gets recreated when the world starts up again
21:26 hmmmm I think what I might do is add in sanity checks for noise values on initialization, but anything after that is up to the user
21:26 celeron55 tiemay: if the world uses some other game than minetest_game, it won't know what game it uses
21:26 celeron55 -> it breaks
21:27 celeron55 it defaults to minetest_game due to legacy purposes
21:27 celeron55 which makes it work for your case
21:27 celeron55 try eg. nodetopia
21:28 celeron55 also if you use some other backend than sqlite3, it breaks
21:29 celeron55 currently you can somewhat safely clear each line from world.mt except the "gameid = <something>" one and the "backend = <something>" one, but more will come in the future
21:29 ShadowNinja hmmmm: No mods are enabled.
21:30 tiemay celeron55: ah, I see now. I'm not sure if bleachbit can erase just part of a file... perhaps the mod enable/disable settings could be stored in it's own file?
21:30 hmmmm can you check out your map_meta.txt?
21:31 celeron55 tiemay: just remove the whole option; you can't do it safely
21:31 celeron55 tiemay: also if you remove the mod options, everything in the world that is made by the mod breaks
21:31 celeron55 the whole content of the file is just inherent to the world
21:32 celeron55 change it to a full-blown "clear worlds" option if you insist on removing that file
21:32 ShadowNinja hmmmm: http://pastebin.ubuntu.com/7252041/
21:32 celeron55 this is just dumb
21:32 hmmmm wow
21:32 hmmmm those are interesting values
21:32 hmmmm why are the last two parameters nuts like that
21:32 celeron55 what are the 3rd and 2nd last parameters?
21:33 hmmmm octave and seed offset, i forget in which order
21:33 tiemay celeron55: doesn't everything in a world from a mod already break after removing the mod?
21:33 celeron55 that's an insane octave either way
21:33 hmmmm yep
21:33 celeron55 tiemay: same thing
21:34 celeron55 tiemay: i'm not going to argue more about this; you know my view on it now and i probably won't ever even try bleachbit if it's policies are like this
21:34 hmmmm hmmmmm and that's the dilemma
21:34 hmmmm do I limit the number of octaves
21:34 tiemay celeron55: dude, chill...
21:34 celeron55 no, i don't chill when people are dumb
21:35 tiemay celeron55: stop pissing on your new contributors, mistakes are part of learning
21:35 hmmmm instead of limiting the number of octaves, i think a better approach would be to bypass the noise lattice buffer and use the slow perlin noise functions to populate the noise
21:36 hmmmm that way people can have 55 octaves if they really want it, and it'll work for everybody, there's just no telling how slow it might be ;)
21:36 celeron55 hmmmm: lol
21:36 hmmmm tiemay, if you can get through the hazing stage I promise everything will be good
21:36 celeron55 33 octaves would take like an hour
21:36 hmmmm eh
21:36 hmmmm more like 2-3 seconds on a fairly modern processor i'd assume
21:37 hmmmm still ridiculous
21:37 celeron55 hmm... isn't it like 2^(33-3) more expensive than a 3 octave noise?
21:37 tiemay hmmmm: hazing is not ok with me, ever...
21:37 celeron55 which is 1073741824
21:37 * hmmmm scratches head
21:37 celeron55 that's probably more than an hour 8))
21:38 hmmmm it should only be 33 iterations per call
21:38 hmmmm the thing that's exponentially decaying is the spread factor
21:39 tiemay left #minetest-dev
21:40 celeron55 so how does the noise thing even iterate
21:41 celeron55 ah yes, it's just linearly increasing the iterations
21:41 hmmmm i'm just talking about the naive point polling functions
21:43 celeron55 some implementation might get that slow though
21:44 celeron55 but the sampling resolution of any octave stays exactly the same in this... which makes the octaves just completely random noise (nothing like perlin) when they go smaller than the sampling resolution of value per node
21:44 celeron55 you could probably convert the remaining octaves to a single noise2d call 8)
21:45 celeron55 (which wouldn't give the same result but would modify the result in the same nonsense way)
21:47 celeron55 i guess it's called aliasing, actually
21:48 celeron55 ...
21:48 celeron55 so where did the values come from in the first place?
21:48 celeron55 the main issue doesn't seem to be how the values are handled but rather the weird values
21:50 celeron55 <+ShadowNinja> celeron55: Did you see my SCons minetestmapper patch?
21:50 celeron55 where is this?
22:32 Eater4 joined #minetest-dev
22:44 hmmmm hmm
22:44 hmmmm https://forums.unrealengine.com/show​thread.php?2574-Why-C-for-Unreal-4&a​mp;p=16252&amp;viewfull=1#post16252
22:45 hmmmm sometimes i wonder if this incredibly enormous lua scripting environment is the best way to do things
22:45 proller and slowest
22:45 proller and for singleplayer only
22:47 restcoser joined #minetest-dev
22:57 Zeitgeist_ joined #minetest-dev
22:57 Zeitgeist_ joined #minetest-dev
23:07 ShadowNinja celeron55: http://ix.io/bEx/diff
23:10 Taoki[laptop] joined #minetest-dev
23:12 ShadowNinja hmmmm: I think that a C++ API using loadable modules (usually shared libraries) would be good.  It would be much less portable though. 23/64 bit and Linux/Windows/OSX versions would need to be projided if they were cross-platform.  Or they'd have to be compiled with it.
23:13 ShadowNinja I don't think that we should drop Lua though, for all of those reasons.  It's also much easier to send to a client and sandbox.
23:15 VanessaE_ ShadowNinja: cue lamefun and his HTML5/javascript rave :P
23:16 ShadowNinja Comments?  http://ix.io/bGm
23:18 ShadowNinja This too, I'll make a pull soon: http://ix.io/bGn
23:33 ShadowNinja ~tell sapier See #1226
23:33 ShadowBot ShadowNinja: O.K.
23:33 ShadowNinja #1226
23:34 ShadowBot https://github.com/minetest/minetest/issues/1226 -- [WIP] Remove dependency on marshal and push the Lua error handler only once by ShadowNinja
23:35 iqualfragile joined #minetest-dev
23:44 kaeza joined #minetest-dev

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