Time  Nick           Message
03:52 RealBadAngel   lookin at logs it was a very silent day yesterday ;)
03:52 RealBadAngel   whats up folks?
04:30 kahrl          have to do some image processing coding for work, not much time for minetest currently :(
04:33 RealBadAngel   on this weekend im moving to a new flat, so i will be offline for a few days
04:35 RealBadAngel   depends on how fast the provider enable the connection for me
04:38 RealBadAngel   before i go offline i will post bumpmapping patch, moon flowers and new dye crafting system
04:41 RealBadAngel   moon flower will provide brand new fluorescent dye
04:42 RealBadAngel   imho way better than torch in crafting recipes to get glowing nodes
04:42 RealBadAngel   and mortar+pestle item to craft dyes
04:44 RealBadAngel   and to PilzAdam, yes im aware minecraft doesnt do this that way
04:45 RealBadAngel   imho we have to stop following mc and go our own way
04:46 RealBadAngel   we are already at this point where mc can look at us at copy our ideas. its Jeb's work
04:47 RealBadAngel   *and
04:49 kaeza          nice idea about flurescent dyes
04:49 kaeza          for example, homedecor lamps (previously from 3dforniture) used torches along with paper
04:50 kaeza          *table lamps
04:50 kaeza          that would invite trouble IRL :P
04:52 RealBadAngel   hehe indeed
04:53 RealBadAngel   crafting burnin torch with something else? nothin but fire
04:55 RealBadAngel   fluorescent dye could be used for crafting nodes that emits light, mesecons displays, monitors etc
04:55 RealBadAngel   and will be logical
04:59 RealBadAngel   and when we get hardware light workin, lamp + fluorescent dye + coloured dye
05:00 RealBadAngel   to get coloured light source
05:15 hmmmm          hmmmmmm
05:15 hmmmm          not a minetest related question but, does anybody know what the GNU make equivalent to bsd's .ifmake is?
05:37 RealBadAngel   no fuckin idea what youre talkin about
05:38 kahrl          not sure either but maybe this? http://www.gnu.org/software/make/manual/html_node/Target_002dspecific.html
05:50 hmmmm          ahh
05:51 hmmmm          thank you actually that does the trick nicely
05:52 RealBadAngel   hi hmmmm
05:52 hmmmm          hello
05:53 RealBadAngel   i will be out for a few days
05:53 RealBadAngel   starting tommorow
05:53 RealBadAngel   im moving to a new flat
05:53 hmmmm          ah nice
05:54 RealBadAngel   i will be back as soon as provider connect me in
05:54 hmmmm          okay
05:55 RealBadAngel   they told me "a few days" but im not sure
08:14 kahrl          any objections? https://github.com/minetest/minetest/pull/742
08:15 kahrl          I've seen this bug happen randomly before and didn't know what caused it
08:15 kahrl          and I don't think anything could break from initializing SEvent structs properly
08:16 sapier1        why is keyInput moved?
08:17 sapier1        keyInput.key to be exact
08:17 sapier1        ok ok its just resorting for better readability
08:17 kahrl          yes, for code symmetry I assume
08:18 sapier1        hmm only a single change of that commit will remain after formspec menu merge
08:19 kahrl          ah, I wasn't sure if that collided anywhere
08:19 kahrl          good that you checked
08:20 sapier1        guiCreateWorld is removed completely guiFormspec is changed to huge extent ... guiTextInput ... wasn't that the part PlizAdam wanted to remove because it isn't used anymore?
08:20 sapier1        but I'm not sure about later thing
08:21 kahrl          last time I checked guiTextInputMenu was still used in your branch for chat input?
08:22 kahrl          if it can be replaced with a formspec that could be done
08:24 sapier1        I didn't do any changes on ingame menus I guess the chat menu could completely be replaced by lua implementation
08:25 sapier1        ok somehow the "/" key needs to trigger lua action .... but if this isn't working already I guess lots of moders will be thankfull to the one adding on_key_press(<any key>)
08:27 kahrl          that would make opening the chat menu a lot laggier though
08:27 kahrl          not sure how annoying that would be
08:27 sapier1        so if you cherry pick the guiTextInput changes this merge will be ok :-) everything else is useless :-)
08:28 sapier1        I add the guiFormspec fix to my patch
08:28 kahrl          how close is the next gen main menu to being ready to merge?
08:29 sapier1        I haven't heared any complains for a while so I guess it's (almost) done
08:30 sapier1        only issues left are "game buttons don't look exactly same as before" and "you have to click once to catch mouse after join"
08:30 sapier1        first is a minor issue
08:31 kahrl          and my dynamic_cast issues, I'll try to hunt them down again today
08:31 sapier1        yes that's the more obscure 3rd one I thought you already fixed it as you didn't say anything about it
08:31 kahrl          I just didn't have time to look at it
08:32 sapier1        quicks solution would be using c casts but I guess that'd hide the real problem
08:32 kahrl          I'd rather not do that
08:33 sapier1        me too as that special case is exactly what dynamic_cast's are for
08:38 kahrl          when I try to run the current next_gen_main_menu branch it fails on startup
08:38 kahrl          http://paste.dy.fi/Soo
08:39 sapier1        curl or non curl mode?
08:39 kahrl          curl is compiled in, not sure if that means curl mode
08:40 sapier1        I don't have curl ;-)
08:40 sapier1        but this needs to be fixed of course
08:45 sapier         can you check latest version? I changed the checks for int json values
08:46 sapier         but I wonder why this did throw an exception for you this means the max_clients value isn't an int value
08:49 kahrl          it works now
08:49 sapier         ok wonder why noone did report this before
08:49 kahrl          I'll make it print out the JSON to see what might be wrong
08:49 kahrl          guess noone checked ;)
08:50 sapier         yes that's probably best explanaition
08:52 kahrl          I peppered mainmenu.lua with print statements and it used to fail in get_favorites("local")
08:52 kahrl          so maybe my local serverlist is somehow broken
08:52 sapier         local? hmm
08:52 sapier         of course there is no max clients
08:53 sapier         ok the previous check was wrong as it checked size of string ... i assumed this'd be 0 in local case
08:53 sapier         obviously a missinterpretation
08:53 sapier         so the new version is fine
08:54 sapier         lol
08:55 sapier         in online case l_get_favourites fetches local list
08:55 sapier         I guess this is exactly inverted
08:56 kahrl          heh indeed
08:56 sapier         not a critical bug but anoying :) fixed
08:59 kahrl          are there any other places where a unexpected online serverlist can crash the client?
08:59 kahrl          such as missing fields or wrong-typed ones
09:01 sapier         I changed the two number values in there to "asInt" and fixed their checks, everything else was there in previous minetest versions too
09:02 kahrl          ok
09:02 sapier         everything else (i know about) are stirng values checked by stringlength
09:02 kahrl          I will audit this part a bit after the new menu is merged and try to fix problems I see
09:03 kahrl          some of it might be security relevant
09:03 sapier         ok thanks ... if you have sparetime try to check the mod and gamemanager too
09:03 sapier         they're disabled by default
09:03 kahrl          not much currently but I'll try to find some time
09:05 sapier         at least modmanager is most likely to be enabled anytime soon as it's intended to be used as central instance for installing mods including moddb online integration (to be done)
09:14 kahrl          I guess I should be happy that you fixed the problem but it makes me feel nervous that I don't understand why it failed ;)
09:14 kahrl          this is the JSON it choked on: http://paste.ubuntu.com/5719324/
09:15 kahrl          all the "clients" and "clients_max" fields seem to be convertible to int
09:18 kahrl          hmm if I read the implementation of isConvertibleTo correctly then isConvertibleTo(intValue) never returns true for a string
09:18 sapier         very strange
09:19 sapier         great
09:19 kahrl          and asInt does not parse a string it just coerces other numeric types
09:19 sapier         ok I guess the solution was to easy .. I'll do it manually and push again
09:19 sapier         what the hell are those functions for if they don't do anything usefull?
09:20 kahrl          they are for giving you a happy fun time developing
09:20 kahrl          until the QA department comes to you and says it doesn't work ;)
09:21 sapier         imho if you supplie a asInt function for a generic datatype this function should be able to give you any convertable value asInt ... like the name says ;-)
09:23 kahrl          why does the online serverlist store some of the numeric fields as string anyway?
09:23 sapier         I don't have any idea :-)
09:23 sapier         most likely because the one writing it didn't bother about datatypes
09:24 kahrl          huh and "damage" is sometimes "1" and sometimes "true"
09:24 kahrl          ah, guess that's directly from minetest.conf
09:25 sapier         I guess I found next bug ... so damage might be 0/false in server list too am I correct?
09:25 kahrl          anything that Setting::getBool can parse, I assume
09:25 kahrl          Settings*
09:26 kahrl          so use is_yes
09:26 sapier         great ... so I need to parse any stupid value ... imho reading is wrong place to fix this ... but it's allmost impossible to revert the errorneous design of serverlist out there
09:27 sapier         there is a is_yes function?
09:27 kahrl          there is, in util/string.h
09:28 sapier         wow :-) good to know I added some helper code to menu to fix this issue ... I'll remove it and use is_yes
09:32 kahrl          is guiLuaApi::isMinetestPath used for sandboxing?
09:33 sapier         partly
09:33 kahrl          I'm asking because it looks like you can circumvent by using a path like /tmp/../etc/passwd
09:33 sapier         sandboxing wasn't main issue of formspec menu but where I needed to check paths I did it in a sane way
09:35 sapier         fs::AbsolutePath fixes that issue
09:36 sapier         a function I've written originaly for security fixes and reused here
09:36 sapier         but
09:36 kahrl          is fs::AbsolutePath used on the path argument? I don't see it
09:37 sapier         yes you're right it's not used for tmp thats an error
09:39 kahrl          I think engine.delete_dir("/home/kahrl/minetest/worlds/../..") would also attempt to delete my home directory
09:40 kahrl          not going to try it ;)
09:40 sapier         no I don't think this would happen
09:41 sapier         absolute path for this is /home/kahrl
09:41 sapier         and isMinetestPath should return false for this
09:44 sapier         ok json as well as isMinetestPath are fixed
09:51 kahrl          ah sorry I didn't see that l_delete_dir already called fs::AbsolutePath
09:51 kahrl          just looked at l_create_dir
09:51 sapier         no problem I didn't see the tmp error at beginning too
09:52 celeron55      it's not secure enough until you can try delete_dir("/home/kahrl/minetest/worlds/../..") without pissing in your pants :P
09:52 sapier         yes create dir doesn't check .. I remember there was a reason for this ... hmm
09:53 kahrl          perhaps fs::AbsolutePath fails if the dir doesn't exist?
09:54 kahrl          yeah probably because of realpath returing ENOENT
09:55 kahrl          returning*
09:57 kahrl          one possibly solution for that is to take fs::AbsolutePath of the parent directory... and if that fails try the parent of the parent, and so on
09:58 kahrl          and call isMinetestPath on the first ancestor where fs::AbsolutePath works
09:59 kahrl          (also fail if the "parent" path component is ..)
10:00 kahrl          actually, not sure if the latter check is needed
10:04 kahrl          ok so the only real issue left is being able to create a dir anywhere
10:04 kahrl          such as engine.create_dir("/home/kahrl/minetest/worlds/../../PWNED")
10:06 kahrl          all the other calls to isMinetestPath use (already used) AbsolutePath
10:06 kahrl          and being able to create arbitrary directories isn't that critical
10:16 sfan5          what about disallowing ".."?
10:18 kahrl          path_share and path_user are ".." in RUN_IN_PLACE builds
10:19 celeron55      i don't see much problem in checking create_dir paths
10:20 celeron55      just strip the last part and see if the result is a legit path
10:22 kahrl          create_dir is like mkdir -p
10:23 kahrl          the parent directory doesn't necessarily exist either
10:23 celeron55      well, then try all alternatives :P
10:23 kahrl          yup, that's what I'm coding ;)
10:27 sapier         I remember reason for not using absolute path in create dir was absolute path failing for non existent paths :-)
10:28 sapier         so creating recursive paths fails
10:29 sapier         if you strip until it works you may miss ".." ... unless you dissallow ".." completely
10:29 celeron55      well, stripping ".." should probably be disallowed (otherwise it
10:29 celeron55      +'s kind of complicated)
10:30 sapier         yes but this will result in a cripled create_dir ... yes it's more secure but to do it that way I need support of some core developers as there will be ppl not accepting that limitation
10:31 celeron55      well, of course it can be done; it basically requires just stripping one legit dir per ..
10:31 celeron55      8)
10:32 sapier         I didn't intend to reimplement realpath ;-)
10:33 sapier         you need to consider "/" "\" mixed pathnames too ... so I guess this is a lot more complicated
10:33 celeron55      hmm, actually all it requires is automatically passing the input of create_dir through some path absolutizer
10:33 * celeron55    is stupid
10:34 sapier         I'm not sure about that but I guess realpath resolves symlinks too
10:34 sapier         yes it does
10:35 sapier         if filling filesystem with dirs would be only security issue left I guess it'd be worth fixing it :-) but there are much more critical errors
10:36 celeron55      well consuming all space or memory or something can't be taken as a concern
10:36 celeron55      it can be done almost always anyway
10:37 celeron55      you can do that even in web browsers :P
10:38 kahrl          hmm CreateAllDirs does something like pos = basepath.rfind(DIR_DELIM_C);
10:38 kahrl          to strip off path components
10:38 kahrl          which ignores '/' on windows although it shouldn't
10:38 sapier         grrr I just realized my last fix for "/tmp" check resulting in inability to create dirs in "/tmp"
10:40 sapier         I guess I need to fix the createdir problem to make this work correctly anyway
10:41 kahrl          sapier, I'm currently doing some stuff related to create_dir
10:41 kahrl          maybe it will already fix the /tmp problem
10:41 sapier         ok if doing so be sure that "isMinetestPath" must not be called with non existent directorys (as it's done atm)
10:42 kahrl          yeah
10:43 kahrl          although, are the calls to fs::AbsolutePath needed outside isMinetestPath?
10:43 kahrl          otherwise I might add an argument "bool mustExist" to isMinetestPath
10:43 sapier         atm it's not needed
10:44 kahrl          and let isMinetestPath handle both situations gracefully
10:45 sapier         this needs rewriting of isMinetestPath ... but I guess it's the way to go
10:46 kahrl          isMinetestPath could also return the absolute path somehow but that might not be needed
10:47 sapier         I'll do the isMinetestPath changes so createdir only needs to call it with false
10:47 kahrl          hang on, I have already started
10:47 sapier         I guess we need a "unrelative" function first removing all ".." from a path
10:48 sapier         from this path you can start stripping elements until realpath succeeds
10:49 sapier         and we can remove fs::absolutepath from delete dir too if isMinetestPath handles this
10:55 kahrl          brb
12:38 kahrl          back
12:54 kahrl          sapier: I wrote RemoveLastPathComponent and RemoveRelativePathComponent functions
12:54 kahrl          https://gist.github.com/kahrl/5684778
12:55 kahrl          I removed DIR_DELIM_C because it is a bad way of checking if a char is a dir delimiter
12:55 kahrl          since in win32 both '/' and '\\' are dir delimiters
12:56 sapier         ok I'm gonna merge this today evening thx
12:57 sapier         I need to fix the is_yes things too I guess it's better not to add a xth variant checking for yes to core if there already is one
13:01 kahrl          sapier: this gist is only the path functions, I've just started working on isMinetestPath again
13:01 sapier         oh ok if you want to do this too I wont stop you
13:01 kahrl          I can try
13:02 sapier         I guess it's good to to the isMinetestPath correct probably at some time it's gona be used to sandbox modding too ... but I doubt this will be anytime soon :-)
13:07 kahrl          hmm I just noticed that getAvailableWorlds checks $path_user/world too
13:07 kahrl          which is not an allowed path in isMinetestPath
13:08 kahrl          so the new menu will probably not be able to delete such a world
13:19 sapier         looks like an error I've not been aware of that location for worlds :-) there are tooo much locations where mintest searches for worlds atm :-). Is this path an old path or a current path?
13:20 kahrl          a very old path but still supported
13:20 sapier         as no world can be created there maybe this is a chance to drop support?
13:21 sapier         I guess nobody would even notice :-)
13:21 celeron55      very old = 0.3
13:21 kahrl          it's easy to add a check for ./world in isMinetestPath
13:21 celeron55      so if anyone updates 0.3 to 0.4 in-place, that's what he would stumble upon
13:22 sapier         is world format still compatible?
13:22 celeron55      yes
13:22 celeron55      i have an automated test utility to check such too
13:23 sapier         what about adding a test on bootup moving worlds to worlds folder?
13:23 sapier         of course asking user if he's ok with it
13:23 sapier         could be done in lua menu too
13:24 kahrl          if it's done in lua you need to add ./world to isMinetestMenu anyway ;)
13:24 kahrl          Path*
13:24 sapier         didn't think about that
13:25 sapier         still imho we need to get rid of some of those old compatibility code
13:25 celeron55      no we don't; it's just a matter of adding a path to a list anyway
13:26 sapier         I'd have to look where this path is used too but I guess there need to be other modifications too
13:26 celeron55      i sure hope this does not interfere at all with any server stuff
13:27 celeron55      and the ability of MT to run a world from any location generally
13:27 kahrl          sapier: lua should not care where a world is located
13:27 celeron55      actually, i *require* this
13:27 celeron55      not just hope
13:28 kahrl          sapier: it gets the path from get_worlds and can assume isMinetestPath accepts the path
13:28 sapier         it's only about mainmenu ... I already realized mainmenu will never fit any requirements of any ppl around. imho it should only implement a sane subset
13:28 celeron55      of course the menu can be unable to list such worlds, but the game should work
13:29 celeron55      people may want to eg. associate minetest with world.mt files and run worlds that way
13:29 sapier         if it's that simple kahrl I don't have any problem with it I'm just worried some things might not work for worlds stored in world and ppl start complaining "but I can delet it why can't I do that"
13:30 sapier         celeron we're only talking about delete not working for worlds stored in "world"
13:31 sapier         at leat atm :-) maybe other things fail too
13:32 kahrl          if they fail they can be fixed
13:32 sapier         yes at cost of adding more complexity to mainmenu ... as always
13:33 kahrl          compatibility with old worlds is important
13:33 sapier         imho this is not a compatibility usecase
13:34 sapier         imho adding a import function keeping base code clean and maintainable is a better way to keep compatibility
13:34 kahrl          well if a certain features works with newly created worlds but not with ones created with an earlier version I'd say that's compatibility
13:35 kahrl          feature*
13:35 sapier         but still let's try if it's ok with just adding this path
13:35 sapier         maybe marking those worlds in world list with a different color
13:36 kahrl          the world will be called "Old World"
13:36 kahrl          getAvailableWorlds does this
13:36 sapier         ok good enough
13:37 sapier         sorry but I've to go now
14:54 kahrl          ohh
14:54 kahrl          I completely interpreted wrong why Old World can't be deleted
14:54 kahrl          the real reason is that mainmenu.lua:568 has a < where there should be a <=
14:55 kahrl          well I fixed isMinetestPath anyway which is good
14:58 kahrl          well mainmenu.lua:557 actually
15:04 kahrl          all my modifications: https://gist.github.com/kahrl/5685589
15:05 kahrl          the map.cpp change is the most dangerous one but I'm fairly certain it is safe
15:05 kahrl          I will also add unit tests for the new path functions I wrote
15:07 celeron55      by the way, apparently GetTempPath can sometimes return an unusable path: http://stackoverflow.com/questions/2365307/what-determines-the-return-value-of-path-gettemppath
15:08 celeron55      (welcome to windows)
15:09 celeron55      i guess it's not like anyone would expect that a modern OS could reliably spit out a path to store temporary files in 8)
15:13 Zeg9           Modern OS ? Reliability ?
15:13 Calinou        you mean, "Modern OS ?"
15:13 Calinou        that's enough
15:17 kahrl          the reason TempPath is needed at all is for unzipping zip files
15:17 kahrl          maybe it can be changed to unzip in place
15:23 kahrl          or use cache/tmp maybe
15:25 Calinou        https://community.dev.fedoraproject.org/packages/minetest (link to "fedora" on downloads page) => 404
15:26 kahrl          Calinou: 502 here
15:27 Jordach        http://i.imgur.com/7ufOThB.png
15:28 kahrl          Jordach: certificate is for *.fedoraproject.org
15:29 kahrl          http://en.wikipedia.org/wiki/Wildcard_certificate "Only a single level of subdomain matching is supported."
15:29 Calinou        I just click "proceed" when I see that kind of warning :phoronix rolleyes:
15:34 celeron55      that mostly doesn't matter as long as you consider browsing a https site without a proper certificate equal to http
15:35 celeron55      which mostly means that if some intelligent person is tracking you, he'll know everything you input or view on the site
15:38 celeron55      (altough that means a lot in the light that some people consider http not worth using for any purpose anymore)
15:38 Calinou        http is much faster than https
15:39 celeron55      based on what
15:39 Calinou        personal tests?
15:39 Calinou        I don't have optic fiber *_*
15:39 kahrl          perhaps because in your case http got cached and https not?
15:40 Calinou        maybe
15:40 Calinou        but https still tends to whine when you're not using signing, and that is annoying
15:40 Calinou        that means you have to pay $90 to verisign & co to be considered "trusted"
15:40 celeron55      https has a slightly more heavyweight handshake (because it has a handshake and http has not); after that's done, it's not slower at all (the encryption is very lightweight)
15:41 celeron55      Calinou: website certificates are a different thing to https, altough in practice you stumble upon them on the net
15:41 kahrl          Calinou: I agree, it'd be nice if there was an option to "just encrypt, don't authenticate"
15:41 Calinou        ah ok
15:41 celeron55      browsers are designed to handle certificates like that because it's the most secure way for end users
15:41 celeron55      (that is, they trust only certain signing authorities)
15:43 celeron55      also, yes, http can be cached easily and https cannot, so that'll make a lot of the speed difference most of the time
16:46 sapier         kahrl unzip in place isn't an option as modmanager tries to verify folder name in order to reduce users problems with invalid mod folders .... that's been one of the original reasons doing this thing at all ;-)
16:47 sapier         and of course "in place" would add additional problems e.g. how to ensure no current mod is overwritten by unzip, how to make sure a mod doesn't modify another mod by unzipping a file to its folder ... lots of things, fixable but lot of work that can be avoided by temp paths
16:50 sapier         kahrl  Calinou encryption without authentification is useless as anyone could claim to be the site you connect to capturing your encrypted traffic reencrypting to destination site
16:59 celeron55      yeah; authentication "as the site you previously used at this url" is quite good though
16:59 celeron55      and browsers do support it; they just warn quite loudly at the first time like in Jordach's screenshot
17:00 sapier         yes I read about an addon for firefox checking sites on p2p base too ... roughly if enough ppl think this site/certificate ist trustworthy its assumed to be
17:01 sapier         of course some other parameters are used to distinguish too
17:02 sapier         large amounts of memory have a big problem too ... they hide memory faults
17:08 RealBadAngel   https://github.com/minetest/minetest_game/pull/165
17:08 sapier         :-) moonflower
17:09 RealBadAngel   im makin now the other part, with dyes
17:10 RealBadAngel   and the fluorescent dye of course
17:10 Zeg9           RealBadAngel: can't wait to get this merged
17:11 sapier         I guess we need to establish some sort of voting system
17:11 RealBadAngel   all the code is by Taoki of course
17:11 RealBadAngel   i mean the flower
17:12 Zeg9           sapier: do you mean for settings/kick/mute/ban ?
17:12 RealBadAngel   i just found out this flower (which is awesome) can be source of fluorescent dye
17:12 sapier         no for what mods shall be added to minetest_game
17:12 RealBadAngel   and this can be used to craft glowing nodes
17:13 sapier         we don't need voting in game as everyone uses minetest with trustworthy users as far as I have been told ;-)
17:13 Zeg9           true
17:13 Zeg9           Well. There is the forum poll but IIRC it needs to make a new topic for every poll...
17:14 sapier         not for core addons for game only
17:42 sweetbomber    hmmm: is any way i can peek on your mapgen_v7 fork? what is your github name?
17:42 sweetbomber    i wanted to get up-to-date on that
17:42 celeron55      kwolekr
17:43 celeron55      dunno if he keeps any online version up-to-date though
17:44 celeron55      doesn
17:44 celeron55      +'t look too much like so
17:44 sweetbomber    oh
17:44 celeron55      most of it goes directly to upstream afaik
17:45 celeron55      so what you have is probably reasonably recent
17:57 sweetbomber    the thing is that, for example, i have heard that biomedef's are loaded from a config file, but by inspecting biome.cpp and biome.h i cant find anything related.. it seems to only register the biomes, not load data...
17:58 sweetbomber    anyway, that was just a detail
17:59 sweetbomber    ill structure my ideas based on current release
17:59 celeron55      https://github.com/minetest/minetest/blob/master/src/script/lua_api/luaapi.cpp#L153
17:59 celeron55      this is really all it takes as input, game content wise
18:00 celeron55      it doesn't implement everything (hmmmm's "decoration" stuff will complement it)
18:00 Jordach        whats the decorations?
18:01 sweetbomber    oh! its called from lua! ok, got it
18:01 celeron55      stuff that the map generator places on ground and inside ground
18:01 Jordach        aye.
18:01 celeron55      sweetbomber: yes, all game content is inserted at load-time into the engine by lua
18:01 hmmmm          right, i'll have to finally commit that sometime soon
18:04 proller        any objections?  https://github.com/proller/minetest/commit/5b4c714a6702cbb66a3a5eade1f0e0dadefd6ce2
18:04 hmmmm          hey sweetbomber, if your idea turns out to work correctly, perhaps you could have caves generated in a similar manner so that there's no need for border blocks in a chunk
18:06 sweetbomber    hmmm, in fact, i already have hatched an idea...but ill test it first on matlab
18:07 sweetbomber    lets rewind to the problem:
18:08 sweetbomber    if the PRNG is used to select the initial node/2dcoord of a structure
18:08 sweetbomber    and that structure can propagate to several chunks
18:09 sweetbomber    then when we are building a chunk it is not trivial to know if in that particular chuck a structure gets to be generated there
18:09 sweetbomber    moreover, the same structure must be generated no matter the order in which chunks are generated
18:10 sweetbomber    this can be summarized as a non-causality problem
18:10 sweetbomber    (imagine the trigger to build a structure as the "cause", and the structure as the "effect")
18:11 hmmmm          actually... use noise3d() instead of 2d, since there are overlapping terrain features
18:11 sweetbomber    (ill go to that point latter)
18:11 sweetbomber    as so, from any place the structure is to be built, we need to know that it is going to be built there
18:12 sweetbomber    what i thought was:
18:13 sweetbomber    imagine a surface like those built from using perlin noise as height
18:14 hmmmm          just so you're aware, this is a really big change
18:15 hmmmm          and it will probably degrade generation performance
18:15 sweetbomber    and use a certain threshold, above which it is considered that such section can accomodate structures
18:15 sweetbomber    this would solve the "know if chunk has structure or not"
18:16 sweetbomber    however there is another thing: whatever the chunk-load order, the structure(s) must be the same
18:16 sweetbomber    for that, one common feature must be taken from that section
18:16 hmmmm          you should keep in mind that with this, you'd still have to calculate every point in a much larger surrounding radius
18:16 sweetbomber    and im thinking taking such feature e.g. from the maximum of that function's at that cut
18:17 hmmmm          to be exact, the radius would need to be the largest size in all dimensions that a decoration could possibly be
18:17 sweetbomber    no, the point is exactly that: you dont have to load what you dont need
18:17 sweetbomber    imagine the following:
18:17 sweetbomber    you calculate the value of that function for the current chunk
18:17 sweetbomber    if it is under the threshold, no prob, there will be no structures in that chunk
18:18 hmmmm          are you using perlin noise for this and not plain positional noise?
18:18 sweetbomber    (btw, this threshold could be used to set the propability of generating a given type of structure)
18:18 sweetbomber    i am not sure yet what would be best for this...
18:18 sweetbomber    but continuing:
18:19 sweetbomber    if the function is above the threshold, the "common feature" must be extracted
18:19 sweetbomber    (this common feature is something that must lead to the same value in a given cut)
18:19 sweetbomber    for now, i am thinking of using the local maximum
18:20 sweetbomber    that maximum would be used to feed the structure generator
18:20 hmmmm          well, i guess it doesn't hurt to try to see if you can come up with something based on that
18:20 sweetbomber    that generator would iterate until it gets to nodes inside the current chunk, and then it would effectively commit them
18:20 hmmmm          i think the details are vague because they're not trivial to figure out, and when you do it might become more complicated than you thought it would be originally
18:21 hmmmm          i still think i'm right though.
18:21 sweetbomber    well, i think there are two aspects difficult to assess apriori
18:21 sweetbomber    first, the 2d noise function: what to use
18:22 sweetbomber    first, the 2d noise function: what to use and what its performance
18:22 hmmmm          nevermind the performance aspect
18:22 sweetbomber    second: the "virtual" (outside current chunk) structure generation must be fast
18:22 sweetbomber    or at least able to jump uncessesary steps
18:23 sweetbomber    regarding the use of 2d or 3d noise, i was thinking about the following
18:23 sweetbomber    each biome could have as parameters the input to a gaussian function, which would determine the probability of it occuring at a given height
18:23 hmmmm          if you're considering perlin noise, a large problem might be dealing with the variance of the steepness of slopes
18:24 sweetbomber    i got some matlab perlin-noise generator, and ill test with that
18:25 sweetbomber    (sorry about my matlab fetish, but i prefer to experiment with matlab than with something else)
18:25 sweetbomber    but using the gaussian curve as height probability would ease a lot of things
18:25 sweetbomber    first, each biome only require two more parameters: variance and mean
18:26 sweetbomber    and it would simplify e.g. the generation of special floating islands (like hidden paradizes, with hidden treasures)
18:27 sweetbomber    or the generation of stuff very deep underground (instead of nether-portals like in mc, it could be possible to dig a lot to the bottom and reach hell
18:27 sweetbomber    what do you think about these ideas?
18:28 sweetbomber    if you have nothing to note, ill try it and pay special attention to your warnings
18:28 hmmmm          pretty complicated and way out there, just don't be disappointed if something doesn't work
18:29 celeron55      i think what you explained in detail about figuring out which chunk contains which structures doesn't really solve anything, or alternatively you're bad at explaining or i am bad at understanding
18:30 hmmmm          i understand what he means, i could probably demonstrate it better by drawing a picture but i couldn't be arsed to do that
18:30 celeron55      probably everyone in here agrees about height-specific biomes though
18:30 sweetbomber    probably i explained myself wrong
18:31 hmmmm          again, if it works out, that's fantastic
18:31 sweetbomber    the height i was talking about was not of the game map, but of a f(x,z) function independent to the height
18:31 proller        will commit !  https://github.com/proller/minetest/commit/5b4c714a6702cbb66a3a5eade1f0e0dadefd6ce2
18:31 hmmmm          proller, looks good
18:32 proller        tested ~week on "sky"
18:32 hmmmm          just a minor stylistic complaint though, i wish everybody would just put if (foo) return; on separate lines
18:32 sweetbomber    the structure generation algorithm is the one that would use all the other parameters already used (heat, humidity and height)
18:32 proller        ok
18:32 hmmmm          also you misspelled neighbors
18:32 hmmmm          but it doesn't really matter
18:33 hmmmm          oh sweetbomber, a problem i'm having with the biomes that i wasn't able to solve is that calculating what biome to use is an O(n) operation and i don't like that
18:33 sweetbomber    n is the number of biomes, i presume?
18:33 hmmmm          yes
18:34 sweetbomber    i believe it happens because it is searched what biome is more close to the head/humm/height of the area?
18:34 PilzAdam       proller, we have discussed to remove the finite liquid checkbox from the GUI some time ago
18:35 hmmmm          yes, finding the least distance
18:35 PilzAdam       proller, it just doesnt work so good and many people enable it without knowing what it is
18:35 sweetbomber    well, at a first glance, search algoritms can be optimized to something like log2(n)
18:35 proller        PilzAdam, and make it default? ;)
18:35 sweetbomber    but maybe it can be simplified
18:36 sweetbomber    because such optimization is based on tree search
18:36 hmmmm          if you could reduce this to nothing more than a searching problem, that'd be a big step in the right direction
18:36 PilzAdam       proller, so, what do you think about it?
18:36 hmmmm          it's not that simple though, since you'd need to calculate all of the distances from each of the biomes
18:37 hmmmm          which is the inherently O(n) part
18:37 celeron55      it's highly likely not an unknown problem in mathematics though
18:37 proller        PilzAdam, many peoples like it
18:37 celeron55      so there are some kinds of optimizations for it for sure (involving precalculating some kind of values, most likely)
18:37 PilzAdam       proller, many core devs dont like it
18:37 sweetbomber    celeron55, you are right
18:37 proller        PilzAdam, i use it without checkbox, remove if you want
18:37 PilzAdam       compared to the default behaviour the finite liquids suck
18:38 proller        default liquid - not liquid
18:38 sweetbomber    i think the biomes could be clustered offline
18:38 Jordach        with some work, this could be added into default: http://forum.minetest.net/viewtopic.php?id=5429
18:38 proller        its blue blocks
18:38 Jordach        and use the bone blocks to craft bonemeal
18:38 sweetbomber    and then, the distance can be calculated first by looking for the nearest cluster,
18:38 PilzAdam       Jordach, we already have regrowing flowers and grass
18:39 sweetbomber    and then only look among the biomes in that cluster
18:39 Jordach        just 2c
18:39 PilzAdam       proller, the good thing in games it that we can define what a liquid is :-)
18:39 hmmmm          sweetbomber, i considered breaking the ranges for heat and humidity and elevation into like, 64 discrete values
18:39 PilzAdam       proller, also, finite liquids isnt really more realistic, water doesnt disappear
18:39 sweetbomber    depending on the number of clusters and biomes per cluster, this would do something inferior to O(n)
18:40 hmmmm          but 64x64x64 is 256 kb
18:40 hmmmm          and not only that but
18:40 hmmmm          you'd lose the precision needed for the elevation part
18:40 sweetbomber    yes
18:40 RealBadAngel   finite liquids require some more work just to become better
18:40 sweetbomber    did you understand my clustering idea?
18:40 PilzAdam       RealBadAngel, yes
18:40 hmmmm          let me re-read that
18:40 celeron55      could some kind of quad tree kind of thing work
18:40 RealBadAngel   but considering removing them at all is wrong
18:40 celeron55      or umm...
18:41 PilzAdam       RealBadAngel, but nobody seems to be interested in that (except proller and he thinks its good already)
18:41 RealBadAngel   lets talk what shall be done to make it better
18:41 celeron55      was the point that you really only need some of the closest biomes for calculating which biome is actually applied?
18:41 celeron55      (usually)
18:41 PilzAdam       RealBadAngel, I could make a list
18:41 RealBadAngel   ideas, we need ideas what shall be done
18:41 hmmmm          you're clustering them based on height, right?
18:41 proller        i think its better than default, and it need more wrk, as whole game
18:41 RealBadAngel   go ahead, make it
18:42 hmmmm          since that's the thing that needs to be exact
18:42 RealBadAngel   proller, ofc, thats the point
18:42 sweetbomber    imagine the simplest case: biomes are only defined by heat and humidity
18:42 proller        last idea - to make water dropable as sand
18:42 hmmmm          that's a simple 2d array
18:42 proller        but not finished
18:42 hmmmm          which wouldn't be a problem to precalculate at all
18:42 sweetbomber    this way, we can pin out the biomes in a plane and its easier to understand the idea (its easily scalable to 3d)
18:42 RealBadAngel   proller, PilzAdam maybe you two can make a list together
18:43 proller        and now i prepare commit to make liquid a bit better
18:43 sweetbomber    then, make clusters from the near biomes
18:43 sweetbomber    so, groups of similar biomes are grouped in one cluster, and so on
18:43 RealBadAngel   u both represent different points of view at this problem, so an agreement between u two shall be a good solution
18:43 sweetbomber    (there are efficient methods to do that, k-means for example)
18:44 sweetbomber    in the end, we have, lets say, 3 groups of biomes
18:44 sweetbomber    and each group (or cluster) has 3 biomes within
18:44 sweetbomber    this is all done offline
18:45 sweetbomber    now, to see what is the nearest biome, given heat and humidity
18:45 sweetbomber    we first see what is the next cluster
18:45 sweetbomber    O(3)
18:45 PilzAdam       RealBadAngel, proller, https://gist.github.com/PilzAdam/5687050
18:45 sweetbomber    and then within that cluster, we only need to search on 3 biomes
18:45 sweetbomber    O(3) again in this cse
18:45 hmmmm          well at this point it becomes a storage problem, not a complexity problem
18:45 RealBadAngel   btw, about biomes, any chance to define gravity for a biome?
18:46 hmmmm          realbadangel, sure, that'll go in with the sky color and other client-side things i guess.
18:46 sweetbomber    well, returning to 3d, a cluster would only be a 3d point
18:46 RealBadAngel   u know what i mean, outer space biomes for example
18:46 sweetbomber    so its quite compact
18:47 Jordach        biomes should also appear at certain heights as well as depths
18:47 proller        PilzAdam, 4 - its generate water filled caves, or dont join caves with ocean
18:47 sweetbomber    outer space like what? i am talking about the biome definitions, not where they should fit in the map
18:47 PilzAdam       proller, whatever fixes the problem
18:48 proller        but now by default ocean restor 0 level
18:51 celeron55      sweetbomber: if you can make a reference implementation of optimized biome calculation, i figure it'd be very useful
18:51 sweetbomber    celeron55: ill give it a shot!
18:52 sweetbomber    hmmmm: got the idea?
18:52 RealBadAngel   sweetbomber, definin gravity for a biome will allow creating near core biomes with high gravity, aether biomes with antigravity, or space colonies without gravity at all
18:52 celeron55      the input to it is biomes in n-dimensional space and later it is asked the closest biome to a point in the space
18:52 hmmmm          sure
18:53 celeron55      n being... umm... 3?
18:53 sweetbomber    omg: i completely mixed up the chat
18:53 celeron55      theoretically anything though
18:54 celeron55      sweetbomber: nah, these other people mix it up
18:54 RealBadAngel   hehe
18:54 sweetbomber    for a moment i tought that RealBadAngel was hmmmm
18:54 RealBadAngel   i just throw in some random ideas ;)
18:54 sweetbomber    and answered as if it has anything to do with what we were talking about :p
18:55 sweetbomber    celeron55: are there any ideas about nether-like world?
18:55 sweetbomber    *about any
18:55 celeron55      depends on who you ask :P
18:56 sweetbomber    lol, i mean as a long-term goal
18:56 PilzAdam       sweetbomber, there is a nether mod ;-)
18:56 RealBadAngel   PA have already a vision
18:57 RealBadAngel   hehe
18:57 sweetbomber    you know, there are things that people could just want to avoid, because it would look like too much like mc...
18:57 sweetbomber    so i like to be cautious about that
18:57 celeron55      it's really hard to see who is even doing decisions here these days, but PilzAdam is used to making final decisions on game content these days
18:57 celeron55      sweetbomber: for sure it shouldn't be the exact same
18:58 sweetbomber    the only thing i tought about different worlds, was to make it more physical: instead of magic portals, why not be possible to go there just by digging or making a giant tower?
18:59 RealBadAngel   lets make a topic asking for ideas
18:59 RealBadAngel   and collect best of them
18:59 celeron55      RealBadAngel: is that a joke? :-D
18:59 RealBadAngel   no, not at all
18:59 PilzAdam       sweetbomber, thats exactly what you can do in my nether mod ;-)
18:59 hmmmm          everybody forgot about the alternative generateTerrain()s =/
18:59 sweetbomber    nice, got to try it
18:59 celeron55      just browse the "feature discussion" section on the forum, you'll have year's worth of ideas for nether there :D
18:59 RealBadAngel   ofc most of ideas will be worth shit
18:59 hmmmm          nether is supposed to occur at -4000 and aether is supposed to occur at 4000
19:00 celeron55      hmmmm: supposed to according to what?
19:00 hmmmm          nether will be nothing more complicated than negative 3d noise, and the opposite with aether
19:00 RealBadAngel   but surely there will pop up some good ones
19:00 hmmmm          my original plans :/
19:00 celeron55      4000 is ridiculous from a gameplay perspective
19:00 hmmmm          remember, we decided against the alternative realms idea?
19:01 celeron55      more like 800
19:01 hmmmm          err
19:01 hmmmm          the intent was to have portals for this
19:01 hmmmm          being able to dig down into it is just a neat side-effect
19:01 sweetbomber    4k  would make that a nice challenge :)
19:01 celeron55      oh whatever; i just hope someone has a clear idea of what they are doing and by some magic chance it will also be good
19:01 RealBadAngel   heh i wanted Portal name for inter server connections ;)
19:01 proller        dig 4k not a big problem
19:01 sweetbomber    loads of cobble need to be digged - clearly a proof of work
19:02 hmmmm          having the alternative realms would require modifying practically all of the Map::getNode/getBlock-type functions
19:02 proller        4k tower - much harder
19:02 sweetbomber    oh! and there is a good thing: after digging 4k for going to nether, the next step would be to spend those 4k nodes in building the tower, lol :)
19:02 hmmmm          everybody thinks that digging seems to be the way to go
19:03 RealBadAngel   possible way
19:03 hmmmm          it's only possible because it's the easiest way to implement it
19:03 RealBadAngel   hard one, but still working
19:03 hmmmm          portals were intended from the start
19:04 celeron55      hmmmm: the original point of sweetbomber is that nether could be just an another kind of cave
19:04 PilzAdam       well, the easiest way is just remove the mapgen part of my nether mod and use it
19:04 RealBadAngel   imho digging your way to helll shall also be possible (and not eliminated)
19:04 hmmmm          well perhaps so
19:04 hmmmm          but i don't like it
19:04 hmmmm          the nether should be negative 3d noise like in minecraft
19:05 RealBadAngel   imagine stairways to hell, or going there with cart
19:05 RealBadAngel   mc CANT do that
19:05 PilzAdam       hmmmm, what will be under the nether?
19:05 RealBadAngel   core
19:05 hmmmm          more nether
19:05 RealBadAngel   no
19:06 RealBadAngel   gravity that will not allow player to move
19:06 celeron55      well, i think there could be much more different levels than only one nether
19:06 RealBadAngel   too high
19:06 hmmmm          you can't put a liquid underneath that or else it'll constantly try to transform and update the lighting, causing Server to become unresponsive
19:06 hmmmm          so perhaps at the very bottom there's bedrock
19:06 RealBadAngel   use physics to make the border
19:06 RealBadAngel   instead of some stupid bedrock
19:06 hmmmm          i'm trying to change as little as possible to other non-mapgen things
19:07 RealBadAngel   make a gravity trap
19:07 PilzAdam       hmmmm, you want a MC like nether; so thats maybe 100-200 nodes high; you want to have the full area under -4200 full of bedrock?
19:07 hmmmm          pilzadam, no, just bedrock starting at -30000 or something
19:08 proller        -30890
19:08 PilzAdam       the MC nether is just one big cave
19:08 hmmmm          well, the exact number at which map generation cuts off depends on the chunksize setting
19:08 PilzAdam       do you want to have several of these caves?
19:08 hmmmm          probably, yeah.
19:08 hmmmm          we'll have different levels of nether
19:09 proller        and increasing lava -30000 -> -30890  (like 100% lava at -30800)
19:09 sweetbomber    well, on thing is sure, something must be done near the coordinates' overflow
19:10 PilzAdam       I like how its currently: just content_ignore
19:10 sweetbomber    i heard (i guess in the forum) something about bending the coord. system
19:10 hmmmm          no idea about that
19:10 hmmmm          no idea how it could even be done
19:10 hmmmm          and obviously it can't be done on the y axis even if it were possible
19:10 sweetbomber    and that would have a nice effect: digging a hole very low would take us to the sky!
19:11 celeron55      falling out from 31000 would be quite a choppy experience
19:11 hmmmm          enough with the chatter... too much talk, not enough doing
19:11 PilzAdam       assume one of the nether caves is 500 nodes high, that would make 52 caves from -4000 to -30000
19:11 sweetbomber    you're right
19:11 hmmmm          all these ideas flying around are worthless unless they're actually put into code
19:11 PilzAdam       thats definetly not good
19:11 hmmmm          let's first have the nether and then we'll decide on the fine details later
19:11 celeron55      hmmmm: their actual viability of being fun is also completely unknown
19:12 celeron55      (until tried)
19:12 hmmmm          right
19:12 hmmmm          well
19:12 PilzAdam       we could also step away from the MC like nether and make it how I do it in my nether mod
19:12 hmmmm          honestly nothing is much fun without mobs
19:12 PilzAdam       just the normal cave system with netherrack that is really hard to break
19:12 hmmmm          pilzadam, that's not very fun
19:12 PilzAdam       have you tried it?
19:13 celeron55      someone get drunk or something and just put mobs into upstream
19:13 celeron55      i
19:13 celeron55      i'm starting to see that the only way 8D
19:13 hmmmm          i haven't played minetest in about 6 months pilz
19:13 Jordach        put simple mobs in master
19:13 hmmmm          pilzadam rather
19:13 Jordach        EVERYONE WILL BE HAPPY
19:13 hmmmm          let's forget about the mobs right now
19:13 hmmmm          let's just work on what we're currently working on
19:14 hmmmm          getting too far ahead of things and it's distracting
19:14 celeron55      hmmmm: there are enough devs in here for one being able to focus on eg. getting mobs done while you work on whatever mapgen stuff you do
19:15 hmmmm          decorations are useful for mob placement
19:15 celeron55      just don't attempt to try to understand everything that is going on
19:15 sweetbomber    want to know my idea of mobs: if we had infinite computer resources they should not exist. They should just be bots of players, where some could have wool, or leather, etc...
19:15 sweetbomber    and oc, could breed
19:15 RealBadAngel   PilzAdam, i answered your post about flower being useless
19:17 sweetbomber    hmmmm: returning to serious business: where can i find your biome-search code?
19:17 sweetbomber    ill pick that first to try my idea
19:17 hmmmm          BiomeDefManager::getBiome()
19:17 RealBadAngel   what others think about moon flower and fluorescent dye in default game?
19:17 sweetbomber    oh, its already commited, sorry...i couldve searched for it if i knew it
19:18 RealBadAngel   hmmmm, celeron55 ?
19:19 PilzAdam       RealBadAngel, see github
19:19 hmmmm          realbadangel, i don't really have much of an opinion on it
19:19 hmmmm          sorry
19:19 PilzAdam       it basically bloats the minetest_game with useless stuff
19:20 PilzAdam       if people like it then they can install the mod
19:20 PilzAdam       there is no need to put this into the game
19:20 celeron55      RealBadAngel: i think it's quite a slim idea until proven otherwise by people building some actually interesting mechanic on it
19:20 RealBadAngel   i have the code that uses it
19:20 Jordach        celeron55, it could be used for glowing mese picksa
19:21 Jordach        which are 2x as strong as a mese pick
19:21 RealBadAngel   but for make it work i need the source
19:21 celeron55      well that
19:21 celeron55      +'s useless
19:21 celeron55      and if not useless, boring at least
19:21 PilzAdam       Jordach, you cant make tools much more faster than the current diamond pick
19:21 Jordach        PilzAdam, you can: instant
19:21 sweetbomber    lol
19:21 RealBadAngel   i wanted the flower to be the source of all glowing things
19:22 celeron55      since when has "glowing things" been interesting at all 8D
19:22 RealBadAngel   so flower -> dye -> combine it with other thigs
19:22 PilzAdam       RealBadAngel, we will maybe have glowstone from the nether
19:22 PilzAdam       wich is much more original than a flower
19:22 RealBadAngel   why the fuck you want the root pull to deliver all the possible uses of it??
19:23 RealBadAngel   fuck the glowstone
19:23 PilzAdam       fuck the flower
19:23 RealBadAngel   we are not mojang empleyees
19:23 sweetbomber    err...
19:23 celeron55      :D
19:23 hmmmm          #shit-that-belongs-in-minetest-delta
19:23 RealBadAngel   i wont agree for glowstone at all
19:23 PilzAdam       why do discussions with RealBadAngel always end like this? :-)
19:23 RealBadAngel   because you are too minecrafty
19:23 celeron55      the minetest chaos development scheme (tm)
19:24 celeron55      i think the flower is better than glowstone
19:24 RealBadAngel   we can do something different and you always go mc
19:24 celeron55      because minecraft
19:24 RealBadAngel   thats why i get angry
19:24 celeron55      but really, why not eg. glow mushroom and put it only in caves
19:24 celeron55      or something remotely less random than a glowing flower
19:25 PilzAdam       btw: Pilz == Mushroom in german, so I like the idea :-)
19:25 hmmmm          in order for that glow flower to work right we need to improve lighting, seriously
19:25 RealBadAngel   i made complete new dye system btw
19:25 celeron55      hmmmm: not really
19:25 hmmmm          not really?
19:25 hmmmm          it's completely dull
19:25 RealBadAngel   and wanted to introduce it with parts
19:25 hmmmm          it's hardly a light at all
19:25 RealBadAngel   no cooking, but crafting
19:25 hmmmm          if you want to make a light seem to be small, you need to have it very bright at the center and die out quickly
19:25 PilzAdam       RealBadAngel, why not put the whole thing into one pull request?
19:25 RealBadAngel   mortar + pestle item
19:25 celeron55      it aims at being a usable gameplay mechanic more than something realistic or particularly pretty
19:26 PilzAdam       so we actually see what you plan
19:26 celeron55      and it's good at it
19:26 hmmmm          w/e
19:26 hmmmm          all that really needs to be figured out is unspreadLight(), and we'd be able to have light with a variable amount of spread
19:26 RealBadAngel   PilzAdam, i cannot do such thing. it requires changes to lotsa mods.
19:27 PilzAdam       RealBadAngel, so you want to introduce new way of producing things
19:27 RealBadAngel   i wanted to make it step by step
19:27 RealBadAngel   dyes
19:27 PilzAdam       like MC brewing and enchanting
19:27 celeron55      really eg. dull lighting is just something on top of the underlying stuff; people all the time confuse the dullness of underlying mechanics with the dullness of look
19:27 RealBadAngel   not all the things
19:27 celeron55      the mechanics don't get better by improving the looks
19:27 celeron55      and so doesn't the play experience; people will be just more disappointed
19:27 PilzAdam       a basic element of MT and MC alpha + early beta was that there are only crafting and cooking
19:27 PilzAdam       this limitation makes it so interesting and fun
19:28 RealBadAngel   we dont know how to get something? lets cook it
19:28 RealBadAngel   for christ sake
19:28 PilzAdam       adding new things like brewing or what RBA wants just destroys this feeling
19:28 RealBadAngel   this is not mc
19:28 RealBadAngel   destroy mc feelin
19:28 RealBadAngel   is what i want
19:28 PilzAdam       I dont agree on bloating minetest_game with shiny new stuff
19:29 PilzAdam       RealBadAngel, I see minetest as a fork of MC
19:29 Jordach        PilzAdam, they werent complaining because they got decent updates worth playing
19:29 Jordach        and paying for
19:29 RealBadAngel   seriously? are you on mojang's pay list or what?
19:30 RealBadAngel   you kill all the pulls that are not mc like
19:30 PilzAdam       adding stuff to not be like MC is ridiculous
19:31 PilzAdam       *with the goal to be
19:31 RealBadAngel   adding stuff that is only mc like is more ridiculous
19:31 celeron55      what i suggest is this: release a short-lived game where people can test more of your game content stuff
19:31 celeron55      so that you can tell based on comments whether it is good or not as a whole
19:31 celeron55      then if it is, it can be merged upstream
19:31 celeron55      if not, then users are glad it wasn't merged
19:32 RealBadAngel   <RealBadAngel> https://github.com/minetest/minetest_game/pull/165
19:32 RealBadAngel   <sapier> :-) moonflower
19:32 RealBadAngel   <RealBadAngel> im makin now the other part, with dyes
19:32 RealBadAngel   <RealBadAngel> and the fluorescent dye of course
19:32 RealBadAngel   <Zeg9> RealBadAngel: can't wait to get this merged
19:32 celeron55      so did zeg9 play it?
19:32 PilzAdam       celeron55, I fear that many people dont think enough to make such decissions
19:32 celeron55      PilzAdam: that's why you need to have a lot of people test it
19:33 celeron55      the problem with MT has always been that we have way too little game design experience to roll our own stuff
19:33 PilzAdam       my experience is that everyone shouts "YESS!!!!"
19:33 celeron55      i included (altough these days i have learned a lot)
19:33 RealBadAngel   and you on the opposite shout "NOOOOO"
19:34 RealBadAngel   i also learned a bit so im not crying out loud to merge the changes
19:34 sweetbomber    if you allow me to express my feelings, i think there should something like a long-term goal and audience
19:34 celeron55      so eg. here's something to read in order to know things game developers generally know: http://gamebalanceconcepts.wordpress.com/
19:34 celeron55      8)
19:34 RealBadAngel   i test the code for weeks before makin pull request
19:34 celeron55      (that's a 10 blog post series)
19:34 sweetbomber    for example, do you think of making MT a game for which kind of people?
19:35 sweetbomber    e.g. redstone in minecraft gave it a geek audience
19:35 sweetbomber    where pvp, brewing and enchanting gave it a fps/rpg audience
19:35 celeron55      sweetbomber: there *definitely* should be a long-term goal and an audience; if i was in lead, i would force there to be one, but i have slipped to other projects
19:36 RealBadAngel   we need them all
19:36 RealBadAngel   its a sandbox
19:37 PilzAdam       I think we cant really find a long term goal
19:37 PilzAdam       there are so many different oppinions out there
19:37 PilzAdam       everyone will just start to cry and shout at the others if his vision isnt accepted
19:38 PilzAdam       we just had a too long time without a real goal
19:38 sweetbomber    RealBadAngel: that would lead to a thing: something like real world. I am not disagreeing, it in fact leads to some interesting things, like doing real business inside MT (much like secondlife)
19:38 RealBadAngel   sandbox means everything
19:38 celeron55      this describes MT's development currently: http://en.wikipedia.org/wiki/Design_by_committee
19:38 RealBadAngel   even things we would like not to allow
19:39 RealBadAngel   thats the freedom of it
19:39 celeron55      "all things" is a thing too; it is very true
19:39 PilzAdam       e.g. if we would have made it clear that MT is a MC alpha clone, then many people who want to merge moon flowers wouldnt even came here
19:39 celeron55      as well as some things by themselves are things
19:39 sweetbomber    but i think that goals must be chosen with some humility, and something like "everything" is awkward..
19:40 RealBadAngel   mt was inspired by mc
19:40 RealBadAngel   but we are not limited to it
19:40 RealBadAngel   theyre suffering atm from lack of ideas
19:40 celeron55      even if someone got a clear vision, it's quite hard to get MT out of the current design state
19:41 PilzAdam       I personally like the idea to make minetest_game a sandbox for mods
19:41 PilzAdam       this means not spamming the whole thing with more and more stuff
19:41 PilzAdam       celeron55, thats what Im talking about
19:42 RealBadAngel   game needs content not only changes to its engine
19:42 celeron55      i don't really know what to do about it, but i want people to understand it
19:42 PilzAdam       RealBadAngel, we definetly do have enough ideas
19:42 celeron55      so they can grab at the chance if it arrives
19:42 PilzAdam       they are just conflicting with each others
19:42 Jordach        1) Tell them to fuck off
19:42 RealBadAngel   PilzAdam, your ideas, no offence, are too minecrafty
19:42 Jordach        (forgot what i had in my head)
19:43 RealBadAngel   all you do, leads to copy mc
19:43 PilzAdam       what does "too minecrafty" mean? is it bad? why?
19:44 RealBadAngel   we all have chosen celeron's project for a reason
19:44 sweetbomber    well, ive to go. theatre at 21:30
19:44 celeron55      RealBadAngel: it is not, in no way, inherently a bad thing
19:44 sweetbomber    bye guys
19:44 celeron55      RealBadAngel: it's only a bad thing if you decide it is a bad thing
19:44 RealBadAngel   ofc its not bad to the bone
19:44 celeron55      RealBadAngel: some people agree with you, some not
19:44 celeron55      RealBadAngel: this is not single-dimensional like you think
19:45 RealBadAngel   it isnt, right
19:45 PilzAdam       I dont want that Minetest turns into a Anti-Minecraft project
19:46 RealBadAngel   me neither
19:46 PilzAdam       where we add moon flowers just to differ from minecrafts glowstone
19:46 celeron55      minecraft is supposed to build upon minecraft alpha; not run away from it or continue with the path MC took
19:46 celeron55      ...
19:46 celeron55      minetest is*
19:46 RealBadAngel   you forced me to say so, in anger
19:46 RealBadAngel   dont use it now please
19:47 RealBadAngel   i just found it out an awesome looking content
19:47 RealBadAngel   and logical source for needed content
19:47 RealBadAngel   and i KNEW you will be against it
19:47 RealBadAngel   because you are always against
19:48 celeron55      i'm not against it; i'm just worried what kind of a larger concept it is a part of
19:48 RealBadAngel   celeron55, not as big as you think
19:48 celeron55      there needs to be a chain of things building towards something in the future, not a random sea of stuff on a blob
19:48 RealBadAngel   just way to get glowing things and more sensible way to get dyes
19:49 celeron55      yes, so it's essentially not wanted
19:49 RealBadAngel   but please notice i splitted it
19:49 celeron55      it's not bad by itself, but in this context merging it doesn't make sense
19:50 RealBadAngel   flower is neutral
19:50 RealBadAngel   dye is a needed content
19:50 celeron55      does someone want to hear my longest standing vision for minetest?
19:50 PilzAdam       yes
19:50 RealBadAngel   dye system i wanted to put on discussion
19:51 celeron55      it's kind of embarrassingly far from most recent development, but not necessarily contradictory
19:51 RealBadAngel   thats why i havent made it one commit, and thx to it PA had a chance to call it useles
19:52 Jordach        celeron55, sure
19:52 RealBadAngel   celeron55, go on
19:56 RealBadAngel   have you said so to stop us arguing? ;)
19:57 celeron55      i'm considering how to actually put it in words
19:57 RealBadAngel   cmon. shoot
20:02 celeron55      what i wanted to do is build upon minecraft's dungeons in some kind of simplistic roguelike fashion with multi-layer dungeons, somehow sometimes making it necessary to build a smallish fortress to defend you from creatures crawling out of some dungeon
20:03 celeron55      it's pretty far from this yay-lets-put-chairs-in-houses thing
20:03 RealBadAngel   as i mentioned before: sanbox
20:03 RealBadAngel   *sandbox
20:04 RealBadAngel   and we may want our fortresses to look good, our way
20:05 RealBadAngel   those things doesnt mean conflict
20:05 celeron55      they mean distraction
20:06 celeron55      people like to focus on some thing
20:06 RealBadAngel   let them allow to focus on what they want to
20:06 PilzAdam       RealBadAngel, "sandbox" is not the solution for everything
20:06 RealBadAngel   not force
20:06 RealBadAngel   freedom of choice is the key
20:07 PilzAdam       a game should give the player some things with limitation, not everything
20:07 RealBadAngel   PA, we are developin a game that from definition shall bring player no limitations
20:08 PilzAdam       limitation is the key why MC is so popular; the player is limited to cubes and has to build with it
20:08 RealBadAngel   but can build ANYTHING
20:08 celeron55      people don't get happy by getting everything in unlimited fashion
20:08 celeron55      nowhere
20:08 PilzAdam       RealBadAngel, no, players cant build circles
20:08 celeron55      not in real life, not in art, not in games
20:09 RealBadAngel   mc like and thus mt are SO GOOD because its a new gender
20:09 RealBadAngel   its not a game anylonger
20:09 RealBadAngel   its a world that let you create
20:10 RealBadAngel   without predefined targets, goals or whatever
20:10 celeron55      one can think it like that; but it doesn't get me interested at all so you need to find somebody else to make it
20:10 celeron55      that's the problem
20:10 RealBadAngel   celeron55, i dont have to look
20:10 PilzAdam       RealBadAngel, I dont see Minetest like this
20:11 RealBadAngel   it already is so
20:11 PilzAdam       if it would be like this then we could stop developing it now :-)
20:11 RealBadAngel   hehe true
20:11 RealBadAngel   its close to it then :)
20:12 PilzAdam       what you mean is is that it has the goal to be like this
20:13 PilzAdam       but this is not a good goal
20:13 PilzAdam       and it wouldnt make a good game
20:13 RealBadAngel   hold on
20:13 RealBadAngel   we do have games system
20:13 celeron55      so RBA's goal really is to just dump anything and whatever somebody cares to create into it?
20:13 RealBadAngel   no no
20:14 RealBadAngel   i havent said so
20:14 celeron55      it could end up being something interesting, but for sure it'd be quite large
20:14 celeron55      we might need to have two distributions, one for that and one for something else
20:14 Jordach        actually two games should be nice: one thats worked on by say, the forum community
20:14 Jordach        and one by the core devs
20:15 RealBadAngel   we tried already with common and minetest_game, partially crossing each other
20:15 PilzAdam       "the forum community" isnt on person and doesnt have one goal
20:15 RealBadAngel   but what about two DIFFERENT games?
20:16 celeron55      RealBadAngel: we already have such ones; they fare fairly well
20:16 celeron55      they're just not officially distributed
20:16 RealBadAngel   may start at one point, but could go different ways
20:16 RealBadAngel   so maybe lets start to do so?
20:16 RealBadAngel   we wont be arguing then
20:16 celeron55      how should it be done
20:17 RealBadAngel   under minetest, lets develop more games just
20:17 celeron55      put everything in one package for windows? how about linux distros? how do we make sure they package the relevant ones
20:17 kahrl          celeron55: that's where sapier new main menu comes in, I'd say
20:17 kahrl          sapier's*
20:17 celeron55      does it allow downloading games from the menu in the future?
20:18 kahrl          not sure if that is planned but I'm sure it could be done
20:18 RealBadAngel   PilzAdam can take care of current minetest_game
20:18 kahrl          lets say if you start the game and it finds only the minimal game, it asks you to choose one of several games and it downloads it automatically
20:18 RealBadAngel   i can take care of some technic one
20:19 celeron55      i think it would be awesome if we wouldn't need to care about distributing anything else on the website to users and distros than the engine, and it would just download stuff from some repository
20:19 RealBadAngel   ofc. engine is separate thingy
20:20 RealBadAngel   but it should allow to browse aviable content
20:20 celeron55      this is in line with the "spirit of 0.4", which is really what we tend to be following still, albeit maybe a bit too vaguely
20:21 celeron55      and it's confusing to think about because it's really not about a game, but about an underlying system
20:22 celeron55      also, this is the kind of programmer porn that one needs to be very careful about; it can sound good even if it really doesn't make any sense from the standpoint of the end result
20:23 RealBadAngel   imho engine shall be developed in manner to allow everything, without goals or limitations
20:24 RealBadAngel   games on the other hand, can be everything
20:24 kahrl          of course it will continue to allow replacing the game, the question is how that capability will be represented
20:24 celeron55      allowing everything means requiring infinite resources
20:24 celeron55      which is unallowable
20:25 celeron55      so we are going to end up with issues trying to figure out what to put in the engine and what not
20:25 RealBadAngel   well, slight limitations then :)
20:26 RealBadAngel   into engine shall go tools
20:27 RealBadAngel   if i. or some1, finds out that for mod to work good, needs something from the engine
20:27 RealBadAngel   thats a good reason to update engine
20:27 RealBadAngel   if it can be done without modyfing engine, it means engine works ok
20:28 RealBadAngel   an example: hardcoded crafting ways
20:28 celeron55      okay so you want to make millimeter-scale microblocks with volumetric lighting and exact physics
20:29 celeron55      you know, we have to have a rule that makes that an immediate "nope"
20:29 RealBadAngel   thats a slight limitation i talked before ;)
20:29 RealBadAngel   ofc we wont allow unrealistic things
20:30 celeron55      but seriously, the engine development goals must be written down
20:30 RealBadAngel   definitely
20:32 RealBadAngel   and seriously too, can i have a copy of minetest_game in minetest project space to take care of, and being different than source?
20:33 RealBadAngel   celeron55, you were timed out?
20:33 celeron55_     i feel i need to begin makin releases again because nobody else seems to understand they need to be made at a reasonably small interval
20:34 celeron55_     +g
20:34 RealBadAngel   have you seen my question celeron55 ?
20:34 celeron55_     0.4.6 is ridiculously old considering how bad it is
20:34 celeron55_     no
20:34 RealBadAngel   and seriously too, can i have a copy of minetest_game in minetest project space to take care of, and being different than source?
20:35 RealBadAngel   ^^ above
20:35 celeron55_     umm wat?
20:35 RealBadAngel   i would like to run a different game
20:35 kahrl          I'm in favor of releasing 0.4.7 soonish
20:35 celeron55_     the question does not make any sense
20:35 celeron55_     kahrl: i'm in favor of releasing 0.4.7 two months ago
20:36 celeron55_     this is ridiculous
20:36 RealBadAngel   it does imho
20:36 RealBadAngel   me and pa, and many have different ideas
20:37 kahrl          well I think a problem was that nobody knew whether the scriptapi move was going to be merged before or after the release
20:37 kahrl          but that is done now
20:37 hmmmm          well now that it's done, we don't need to worry
20:37 RealBadAngel   im not talkin bout common etc like we tried before
20:37 hmmmm          perhaps right here is a good point to release 0.4.7
20:37 RealBadAngel   agree
20:37 hmmmm          nothing's been committed for about 5 days, right?
20:37 hmmmm          there, that's almost a week of feature freeze
20:37 kahrl          there are some pull requests on github that fix things broken by the scriptapi move
20:38 hmmmm          i was afraid of that
20:38 RealBadAngel   i have the code ready to pull
20:38 RealBadAngel   but i think it can wait for next release
20:38 RealBadAngel   i mean bumpmapping
20:38 celeron55_     okay, only fixes allowed then; throw in all the fixes and that's 0.4.7 then
20:39 hmmmm          right
20:39 kahrl          I'd have liked to have the new main menu in but it's much too unstable at the moment
20:39 hmmmm          i need to stop screwing around and finish what i was working on so i can do more minetest
20:39 celeron55_     kahrl: that's a concern that should not exist
20:39 RealBadAngel   better merge it into indev and try it for a few weeks
20:39 celeron55_     kahrl: releases should be common enough for nobody to think like that
20:40 kahrl          well the commit I tried this morning just crashed on startup
20:40 hmmmm          right, these are *patch* version levels we're releasing, not minor versions or anything
20:40 hmmmm          so it's settled
20:40 hmmmm          in 2 days, we'll mark it as 0.4.7 and release it
20:40 RealBadAngel   i thought bumpmapping code was ready, then after i made it public i rewrote it a few times thx to reports
20:40 celeron55_     seriously, even having one 0.4.x per week won't be even close to too often
20:41 kahrl          celeron55_: I think that's what daily builds are for
20:41 hmmmm          celeron wants a rolling release
20:41 hmmmm          i don't though
20:41 celeron55_     it depends on the amount of features put in of course
20:41 celeron55_     and fixes
20:41 RealBadAngel   hmmmm, something in the middle i think
20:41 RealBadAngel   we do have too many changes onboard already
20:41 hmmmm          i also don't want to start pumping out new versions on a set date, that's ridiculous
20:42 kahrl          I like having at least a week of feature freeze for most releases
20:42 hmmmm          by the way, by "freeze" i mean we don't commit anything, but that's not what most other projects do
20:42 RealBadAngel   like hmmm noticed 5 days are frozen already
20:43 RealBadAngel   lets add weekend to it just
20:43 hmmmm          they'd make a stable branch at that point when they think it's good
20:43 hmmmm          the reason why we stop all activity is because honestly, nobody focuses on anything except upstream
20:43 celeron55_     well this is quite a small project really
20:44 kahrl          a feature freeze is pointless if bug fixes can't be committed
20:44 celeron55_     it doesn't make much sense to wiggle with release candidate branches and backporting of stuff
20:44 kahrl          or am I misunderstanding completely?
20:44 hmmmm          erm
20:44 hmmmm          you're misunderstanding, that's called a code freeze
20:45 hmmmm          when i mean don't commit anything, i meant to say no commiting new features
20:45 RealBadAngel   brb
20:45 hmmmm          s/mean/say/
20:46 kahrl          ah ok I agree with that then
20:46 hmmmm          make is such a pile of crap.
20:46 hmmmm          ugh.
20:46 celeron55_     why do you think so
20:47 celeron55_     i think make is pretty good, altough making certain special things with it tends to be unintuitive
20:47 hmmmm          if the whole "tabs and spaces have different meaning" thing isn't enough of a problem, then it's the amount of crypticness
20:47 hmmmm          the amount of variance between BSD and GNU make
20:47 hmmmm          everything is unintuitive
20:47 celeron55_     oh well that BSD/GNU thing might be a problem
20:48 celeron55_     i've never even tried to use BSD make
20:48 hmmmm          ugh, it's like perl at some points
20:48 hmmmm          SRCS = {$@%##SRCS:}$@#?/*"}
20:48 kahrl          why not use cmake or something like that? I mean it's still cryptic but at least has the same syntax everywhere
20:48 hmmmm          see, I've been considering that
20:49 hmmmm          adopting cmake for my own projects
20:49 celeron55_     in smaller projects i tend to use gnu make everywhere
20:49 hmmmm          i'd have to use something like autotools for projects that are likely to be cross-compiled onto some mobile device though
20:49 celeron55_     (which is, on linux and windows)
20:50 hmmmm          oh yeah, there's also NMAKE :/
20:50 celeron55_     cross-compiles are a pain for sure; cmake can sometimes be roughly equal to magic in such cases
20:50 kahrl          mingw works well for small projects so no need to mess with nmake
20:50 celeron55_     yes i use mingw
20:50 hmmmm          well nmake is pretty useless aside from where it's necessary (drivers)
20:51 hmmmm          it's just funny how there's yet another flavor of make
20:52 celeron55_     everyone thinks they're more important than others and that's where it ends up
20:52 celeron55_     it happens in everythin
20:52 celeron55_     +g
20:53 celeron55_     then everyone just hates each other 8)
20:53 kahrl          NIH syndrome, or embrace, extend and extinguish
20:53 Exio           rewrite minetest hmmmm
20:54 hmmmm          i say we nuke it and start all over.
20:54 hmmmm          in java.
20:54 Exio           or better, C#!
20:54 kahrl          whitespace. you said you like significant tabs
20:54 celeron55_     let's make it look like skyrim and be a p2p MMO
20:55 hmmmm          it'll be an AAA game and targeted at the "hardcore gaming" crowd, get some endorsements from mtn dew, razer, and have a soundtrack by Skrillez
20:55 hmmmm          s/skrillez/skrillex/
20:56 celeron55_     by the way, at this last minute, i declare the troll language of the day to be X++
20:56 celeron55_     it's made by microsoft and it's like C#, except not exactly
20:56 celeron55_     8D
20:57 celeron55_     (it's 23:57 here now)
20:57 Exio           utc+?
20:57 Exio           1, 2?
20:57 celeron55_     2+DST
20:57 hmmmm          probably utc+5?
20:58 celeron55_     =3
20:58 celeron55_     "X++ is part of the MorphX development platform that you use to construct accounting and business management systems"
20:58 hmmmm          i remember time zones based off of est
20:58 celeron55_     this probably means it's prefect for gamedev
20:58 kahrl          we need timezones in minetest
20:58 RealBadAngel   back
20:58 Anchakor       hi all :)
20:58 kahrl          anyone who uses /time would have to remember them. It would be totally realistic and awesome!
20:59 Anchakor       kahrl found his way back? nice
20:59 RealBadAngel   kahrl: google for fiatlux.py
20:59 celeron55_     yes and how about pooping
20:59 RealBadAngel   i ported it into engine
21:00 kahrl          celeron55_: well irrlicht doesn't support olfactory devices yet so it wouldn't be the full experience
21:00 hmmmm          but, seriously
21:01 celeron55_     kahrl: maybe it could be shown two times as big on the screen to compensate
21:01 hmmmm          i'd say the reason why I hate something cryptic like make but I love C at the same time is because make is merely a means to an end, you don't really care to spend too much time with make
21:01 celeron55_     are you saying C is some kind of a god
21:01 Anchakor       I see MT got relicensed under LGPL, nice
21:01 hmmmm          C is amazing
21:01 hmmmm          C is the light and the truth
21:02 Anchakor       though it isn't library
21:02 celeron55_     by the way, has someone read about rust here? to me it seems quite awesome as long as the infrastructure around it gets more mature: http://www.rust-lang.org/
21:02 celeron55_     it fixes practically all of the issues i know of C and C++
21:02 hmmmm          i still never figured out why LGPL was chosen
21:02 celeron55_     ...except compile time
21:03 celeron55_     (partly)
21:03 celeron55_     (compared to golang)
21:04 Anchakor       hmmmm: I suggested MPL: http://c55.me/minetest2/wiki/doku.php?id=dev:license_change
21:04 RealBadAngel   so what about me running second game (with technical twist) ?
21:05 RealBadAngel   i wont yell at PA and would expect the same from him ;)
21:06 RealBadAngel   we are thinkin in totally different manners, so the games will be different
21:07 RealBadAngel   you all know that some of us are older
21:07 kahrl          RealBadAngel: why does it have to be under github.com/minetest/?
21:07 celeron55_     RealBadAngel: there's the issue there that modders will hate you for making stuff that makes it impossible to make some mods compatible with both it and minetest_game
21:08 celeron55_     but i guess haters can hate
21:08 RealBadAngel   ofc they will
21:08 RealBadAngel   its their job to do so
21:08 RealBadAngel   but look at the current state
21:09 RealBadAngel   im the developer with ideas i would like to force
21:09 RealBadAngel   but PA says no
21:09 RealBadAngel   gimme just another room
21:09 RealBadAngel   you know im skilled enough
21:10 celeron55_     i don't see any reason for it to be under minetest
21:10 RealBadAngel   i can give you some
21:10 RealBadAngel   redstone - mesecons
21:10 RealBadAngel   enough?
21:10 celeron55_     wtf?
21:11 RealBadAngel   core of minecraft success
21:11 kahrl          people will ask why it is under minetest and some already popular games like minitest and realtest not
21:11 celeron55_     the only reason to put things under github.com/minetest is if they are intended to be developed by the core team
21:12 celeron55_     if not, then there is no reason to and they won't be
21:12 RealBadAngel   im a dev, but cant push to mintetest_game
21:13 RealBadAngel   im representing older part of the community
21:13 RealBadAngel   im 40 yrs old
21:13 Anchakor       ageism!
21:13 RealBadAngel   i like other things than most of you
21:14 RealBadAngel   lol, just the truth
21:14 RealBadAngel   thats why we have those stupid argues from time to time
21:15 RealBadAngel   vanessa, mauvebic, me, and some others who left alredy
21:15 RealBadAngel   we need our space
21:15 kahrl          I don't think there would be no arguments if we only allowed people of age 25 to talk on IRC ;)
21:15 RealBadAngel   hahaha
21:15 RealBadAngel   :P
21:16 iqualfragile   actualy i think that an agelimit at about 14 would be a good idea
21:16 hmmmm          [05:15 PM] <kahrl> I don't think there would be no arguments if we only allowed people of age 25 to talk on IRC ;)
21:16 hmmmm          there wouldn't be much talking at all - what is the probability that there are two people exactly 25 years old on the same irc channel at the same time?
21:16 kahrl          good point :)
21:16 RealBadAngel   i cant see leaving messages
21:17 RealBadAngel   all 25's here?
21:17 hmmmm          i'm 21
21:18 RealBadAngel   so, again, whats wrong in TWO different games...
21:18 * kahrl        is of the wrong age but will continue to talk until hit by the banhammer
21:20 RealBadAngel   also run by TWO different groups of ppl
21:20 RealBadAngel   i mention it because last attempt was run by the very same person
21:20 kahrl          seeing as PilzAdam is pretty much the only one committing to minetest_game right now
21:20 RealBadAngel   why?
21:20 kahrl          I don't think there would be much interest in running a second game by the other devs
21:21 RealBadAngel   because we all know our commits will be rejected
21:21 RealBadAngel   simple fuckin flower is the evidence
21:21 kahrl          so I'd guess you'd be the only one who commits so you can put the repository under your own github account as well
21:22 RealBadAngel   no
21:22 RealBadAngel   youre completely wrong
21:22 RealBadAngel   there are lots of ppl that would like to commit
21:22 RealBadAngel   but they know it makes no sense
21:22 celeron55_     so are those people in this list? https://github.com/minetest?tab=members
21:23 RealBadAngel   many of them left already
21:23 celeron55_     what's your point then?
21:23 celeron55_     you can cry all you want, but it won't make any difference
21:23 kahrl          you can give other people push access to your github repos as well
21:24 RealBadAngel   i could
21:24 RealBadAngel   but thats not the point
21:25 RealBadAngel   celeron55_,  you would call VanessaE an outsider? mauvebic? cornernote?
21:25 Anchakor       what is wrong with running your own repo?
21:25 celeron55_     no? what the fuck man?
21:26 Kray           wtf :D
21:26 RealBadAngel   i just named folks of the similar age
21:26 Kray           all developers having push access to the same repo pretty badly cripples the idea of git
21:26 hmmmm          mauvebic is old?
21:26 RealBadAngel   and notice we all have different ideas than 20's
21:27 RealBadAngel   dont call us old :P
21:27 RealBadAngel   middle age :P
21:28 RealBadAngel   i just want a space for us
21:28 RealBadAngel   not somwhere outside
21:29 celeron55_     hint: you can make organizations on github by yourself too
21:29 RealBadAngel   it is not funny at all celeron
21:30 celeron55_     if you want to have a clear team to work on your game, that's an obvious solution
21:30 RealBadAngel   we all are workin on the very same game
21:30 celeron55_     also, i'm not trying to be funny at all
21:30 hmmmm          i think yall are taking this way too seriously
21:30 RealBadAngel   im trying to find a solution for us all
21:31 hmmmm          i haven't really been listening closely, but the problem is that you want to have your own minetest game?
21:31 hmmmm          and you want it to be an official one
21:31 RealBadAngel   you think its funny to code something and knew in advance that PA will reject it?
21:31 hmmmm          hmmmm
21:32 hmmmm          PA has his own ideas of how minetest_game ought to be, i'll admit
21:32 RealBadAngel   and it is not only my opinion
21:32 hmmmm          some of the changes are things that i wouldn't do myself
21:32 Anchakor       you want your own MT game, but still be distributed with MT and in its git repo?
21:32 RealBadAngel   many folks left the community because of him and his attitude
21:32 celeron55_     minetest's git repo has nothing to with distributing minetest
21:32 hmmmm          that is true
21:33 RealBadAngel   engine is the engine
21:33 RealBadAngel   i say allow multiple official games
21:33 jin_xi         idk, but i find the whole idea of a 'default game' quite unnecessary. it obviously leads to much discussion, friction and frustration
21:33 celeron55_     if somebody didn't notice, this turned into some ridiculous drama just after there was a solution to everything, once RBA somehow forgot it
21:33 celeron55_     jin_xi: i do see it, yes
21:34 hmmmm          most of this has to do with the idea that the engine is necessarily paired to some certain game by default (well, that game is actually minimal, but whatever)
21:34 iqualfragile   jin_xi: a default game is important to have for new players
21:34 celeron55_     jin_xi: but we have no solid means of distributing games for the engine without putting them in by default, so this is really an issue
21:34 iqualfragile   as a starting point
21:34 RealBadAngel   celeron55_, i havent forgot anythin
21:35 hmmmm          if we went with my idea of advertising minetest as its own game and then saying "oh, yeah, you play this using the minetest voxel engine, which you can get here" that might help
21:35 RealBadAngel   even game was splitted into parts before
21:35 celeron55_     so how about focusing all efforts on making the engine able to pull games from some central repo (for newbies) and just simply not distribute it ever anymore with any preinstalled game?
21:35 jin_xi         i think stuff like minitest and such should be promoted and only have something like 'made with minetest' links on their sites.
21:35 iqualfragile   about the direction of minetest_game: i have allready said that but what about publishing an document which clearly states which aims minetest_games wants to reach
21:35 kahrl          not the renaming discussion again :)
21:35 kahrl          it went nowhere
21:35 hmmmm          celeron, yeah, that's something that most people would like
21:35 RealBadAngel   there was only one person who decided what can be merged
21:35 iqualfragile   i think that would remove a lot of the friction
21:35 RealBadAngel   and that was the problem
21:35 celeron55_     if not that way, then the other way is what jin_xi now said
21:35 celeron55_     the "minetest as library" way
21:36 hmmmm          what if minetest_game had a different community as minetest
21:36 celeron55_     iqualfragile: who will do it? everyone knows it would help, but there is no authority on it
21:36 kahrl          celeron55_: the repo thing sounds like a good goal for 0.4.8
21:36 iqualfragile   says celeron55_…
21:36 RealBadAngel   tip: rename it to pilzadams_game
21:36 hmmmm          i do agree though, only one person with with their territory marked over minetest_game isn't very good
21:37 hmmmm          we all have our own territory
21:37 celeron55_     iqualfragile: i don't have any idea what i want from it
21:37 iqualfragile   it seems like PA has
21:37 iqualfragile   what about forcing him to publish such an document
21:37 hmmmm          like mine for example are the mapgens, biome manager, emerge manager, and what not
21:37 iqualfragile   it would be a lot easier to discuss about it then
21:37 hmmmm          but changes to it aren't mostly subjective and user visible
21:37 hmmmm          that's where the problem is
21:37 celeron55_     PilzAdam: write down your plans of minetest_game for the next 2 years and publish it
21:38 RealBadAngel   PilzAdam has taken minetest_game, and if somebody cant see it is either blind or stupid
21:38 hmmmm          the actual problem is how people decide what becomes what in minetest_game
21:38 celeron55_     PilzAdam: how does that sound like? 8)
21:38 kahrl          I wonder what my territory is
21:38 iqualfragile   about the central repository for games: i could add that into mmdb quite easily
21:38 hmmmm          i feel like a dog pissing on source files, haha :(
21:38 iqualfragile   but i would rather wait and see how the modmanaging part works out
21:38 celeron55_     i see kahrl's mission as making minetest 0.4 actually happen somehow like it was originally planned as
21:38 hmmmm          err, skunk rather
21:38 * RealBadAngel usually wanders in mapblock_mesh.cpp
21:38 iqualfragile   celeron55_: PilzAdam is afk
21:39 hmmmm          so there's what we do
21:39 Anchakor       seems like a good idea to not have any official minetest game (except minimal) and distribute with like 5 most popular MT games
21:40 Anchakor       reduces community friction
21:40 hmmmm          either clearly define some democratic way of getting things in minetest_game, OR we add more separation between engine and game
21:40 celeron55_     Anchakor: that's what i am more and more leaning towards
21:40 iqualfragile   i am clearly agains havin none or multiple default games
21:40 celeron55_     Anchakor: or alternatively distribute it with nothing and let the game/mod manager do it
21:40 iqualfragile   as it will lead to a lot of confusion for newbs
21:40 RealBadAngel   which is bad
21:40 hmmmm          iqualfragile, it's better to have multiple rather than none
21:41 celeron55_     the only thing that is newbiew-friendly is "minetest as library"
21:41 celeron55_     but it's not very 0.4y
21:41 RealBadAngel   a few maintained and official games are better
21:41 hmmmm          or at least a "you don't have to read anything" kind of way to set it up if it comes with none
21:41 iqualfragile   this is allways an issue in open-source projects: some people are unhappy with an decision and say: make it an option
21:41 Anchakor       iqualfragile: it's kinda like the good old Half-Life mods, if there is a clear way how to launch one, there is no problem
21:41 iqualfragile   and in the end there are 119919 options and nobody has any idea whats happening anymore
21:42 RealBadAngel   notice you gonna get two solid (and opposite) maintainers
21:42 RealBadAngel   me and Pa
21:42 Anchakor       iqualfragile: you are using a strawman argument
21:42 iqualfragile   i am?
21:42 kahrl          iqualfragile: that's why I'm in favor of a clear, unconfusing screen that shows up when you first run the client and lets you choose a game to download
21:42 hmmmm          we can put CURL to better use this way
21:42 iqualfragile   i would rather distribute one game with minetest
21:43 iqualfragile   one and only one
21:43 Anchakor       why?
21:43 iqualfragile   there should be a clear definition of that games goals
21:43 RealBadAngel   sandbox
21:43 RealBadAngel   we cant
21:43 iqualfragile   and then people should have an easy to use and working mechanism of adding new gamemodes
21:43 celeron55_     iqualfragile: your thing can only happen if there is a leader with a clear plan
21:43 celeron55_     i am not willing to work as such for MT anymore
21:43 celeron55_     it's too painful
21:44 iqualfragile   no need for a single leader
21:44 celeron55_     then you're responsible for figuring out how to do it
21:44 iqualfragile   it would be enought if people would try communicating more
21:44 celeron55_     no it's not
21:44 RealBadAngel   thats a price and valor of community project
21:44 celeron55_     people have so different goals it doesn't work
21:45 kahrl          I would love to see the dungeon crawler game though :)
21:45 iqualfragile   they do? im not shure about that, i just think that people are not communictating well and that increases the percived delta
21:45 Anchakor       iqualfragile: you single game is minetest_minimal, happy? :)
21:45 RealBadAngel   kahrl, why not?
21:45 celeron55_     iqualfragile: if you read the full discussion today, you can see it clearly between RBA and PilzAdam
21:45 jin_xi         i thought about doing a nethack/mt mashup but have not yet implemented anything...
21:45 celeron55_     iqualfragile: their plans for the game are like 90 degrees apart
21:46 iqualfragile   Anchakor: no, the single game needs to be playable, its for new players of minetest
21:46 RealBadAngel   PA runs something similar to mc, i would like to run enhanced one, with technical twist
21:46 RealBadAngel   you can go dungeons
21:46 iqualfragile   stop
21:46 iqualfragile   write that down into an easy accessible document, please
21:47 iqualfragile   perhaps a wiki page
21:47 iqualfragile   inclue wich subsystems you would want to add
21:47 iqualfragile   i am shure that we will find some kind of acceptable solution
21:47 iqualfragile   *sure
21:47 RealBadAngel   in fact theres nothing to do
21:47 RealBadAngel   just make new folder
21:48 RealBadAngel   and make real differnt games with time
21:48 iqualfragile   creating a folder and throwing in some mods is not how you make a gamemode
21:48 RealBadAngel   time
21:48 Anchakor       iqualfragile: I guess you are the naive kind of a guy who thinks people can always work together only if they tried harder to understand each other
21:48 iqualfragile   a gamemode needs to be more then just its mods, in my oppinion at least
21:49 RealBadAngel   of course
21:49 RealBadAngel   but time will effect with changes
21:49 RealBadAngel   that one would accept and other reject
21:49 RealBadAngel   sum of this will produce differnt game
21:50 Anchakor       but you can't work together on going different directions... minetest engine allows many game modes, trying to be impartial to the game projects is good
21:50 iqualfragile   well, then lets define some criteria for a good game
21:50 jin_xi         huh. terminology is not helping us here. cant mod a non existing games... so we should call them components maybe... or plugins even lol. and dont get me started on gamemode, specially in the context of making them more different from each other
21:50 RealBadAngel   we will need the same engine still
21:51 RealBadAngel   but will differ in usage of it
21:51 kahrl          sapier: I added some test cases for the new path functions (only tested in linux so far). Here's all my changes as a diff: https://gist.github.com/kahrl/5688237
21:51 iqualfragile   sapier: whats the status of the modmanager?
21:51 RealBadAngel   im using many parts of code wrote by PilzAdam
21:51 kahrl          the diff also has the new isMinetestPath etc.
21:52 RealBadAngel   shit we are all using Minetest :)
21:52 Anchakor       iqualfragile: if at start you are preseted with 5 game mode choices in order of the community popularity vote, that gives new players a good idea what is the "main" MT game without it being set in stone and undemocratic
21:52 sapier         << reading backlock .... lots of comments since I left
21:52 Anchakor       presented*
21:52 sapier         iqualfragile local mode is completed
21:52 sapier         download only has a button :-)
21:53 iqualfragile   when do you estimate that part will be finished?
21:53 celeron55_     i really would want to put good game mode developers at the front line of development, everything else just following their demands
21:54 celeron55_     if anyone can adjust their work to work like that, please do
21:54 celeron55_     it'll please everyone in the long run
21:54 RealBadAngel   we can get our niches, sure
21:55 kahrl          sapier: in case it got obscured by the thousands of comments it got decided to make a feature freeze and release 0.4.7 after this weekend
21:55 celeron55_     the problem is, who is good; nobody here really has any merits
21:55 celeron55_     someone like the dwarves developer could be counted as such
21:55 sapier         Jordach NOT EVERYONE will be happy with simplemobs ... imho their stupid and buggy ;-P
21:57 RealBadAngel   celeron55_, what i proposed was not a total anarchy in creating many default games. but propably two of them to let OLDER CODERS to code
21:57 kahrl          so yeah I guess I'll go through the pull requests and check those that claim to be bugfixes
21:57 celeron55_     RealBadAngel: stop that already; nobody is interested
21:57 kahrl          also the bug reports
21:58 RealBadAngel   nobody is interested?
21:58 proller        PilzAdam, who dislike finite_liquid except you?
21:58 iqualfragile   kahrl: good idea
21:58 kahrl          https://github.com/minetest/minetest/pull/237
21:58 celeron55_     proller: people dislike it because it doesn't work too well and noobs enable it without knowing
21:58 iqualfragile   i would apreciate it if devs would care more about the pull requests
21:59 kahrl          #237 seems badly implemented according to PA's comment so I will leave it
21:59 iqualfragile   proller: i dislike finite liquids, it does not work for me
22:00 kahrl          also needs a huge rebase probably
22:00 kahrl          https://github.com/minetest/minetest/pull/296 bug fix for huge map and add-on of -xmin -xmax -zmin -zmax param (block unit (x 16 cubes))
22:00 RealBadAngel   well, if nobody is interested then :P
22:01 iqualfragile   kahrl: https://github.com/cyisfor/minetest/commit/45afcec615c35e2bacf3abc7457c791281d9342b#L0R1281
22:01 sapier         I guess yesterday silent day is completely fixed now :-)
22:01 proller        iqualfragile, does not work what?
22:01 sapier         thx kahrl I found it :-)
22:01 iqualfragile   on my computer
22:02 sapier         proller I don't dislike finite liquid but I think it's far from finished too
22:03 sapier         kahrl so formspec mainmenu is for 0.4.8? I guess this gives oportunity to integrate moddb too?
22:03 kahrl          iqualfragile: well I haven't seen the bug it claims to fix, the pause menu does not consume more CPU time than the normal game
22:04 kahrl          sapier: yep
22:04 iqualfragile   kahrl: its just formulated badly
22:04 sapier         still I'd prefere integration as soon as possible once 0.4.7 is released to find the bugs ... I guess scriptapi separation will have some errors after 0.4.7 release due to it's late merge
22:04 iqualfragile   it would just be smarter to reduce the fps when you are looking at an menue anyways
22:05 sapier         did I just miss result of release strategy or wasn't there any ?
22:05 RealBadAngel   it will stay the same
22:05 kahrl          iqualfragile: guess it makes sense on old hardware though just calling device->sleep(const) isn't the way to do it
22:06 RealBadAngel   what pa likes or not
22:06 kahrl          anyway this is not a critical fix so I'll leave for after 0.4.7
22:06 RealBadAngel   with c55 hidin behind pa's back
22:07 sapier         https://github.com/minetest/minetest/pull/741 needs to be merged for 0.4.7
22:08 kahrl          sapier: it does
22:08 iqualfragile   kahrl: https://github.com/minetest/minetest/pull/744 would be nice to have
22:08 kahrl          iqualfragile: that's a feature; it's feature freeze now
22:08 sapier         https://github.com/minetest/minetest/pull/739 this one too
22:08 iqualfragile   i would rather consider the spawining buggy
22:08 sapier         https://github.com/minetest/minetest/pull/740  i guess this one wouldn't be bad too
22:09 celeron55_     stuff can be merged only if they fix critical bugs or fix bugs caused by recent changes
22:09 kahrl          did you check the code of #739 and #740 if they are correct?
22:10 iqualfragile   it stops people from spawning midair, so its a bugfix :p
22:10 kahrl          iqualfragile: has a good chance of introducing new bugs so it'll have to wait in my opinion
22:10 sapier         celeron those commits fix scriptapi separation bugs (except last one) that one fixes a generic bug
22:10 iqualfragile   well, ok
22:11 sapier         I need to reboot my mouse got frozen ... I hate this bug
22:12 kahrl          back to going through them back to front
22:13 kahrl          https://github.com/minetest/minetest/pull/296 not caused by anything recent
22:14 kahrl          https://github.com/minetest/minetest/pull/324 same, also looks controversial
22:14 sapier         did anyone notice the #pull links in irclog is still pointing to https://github.com/celeron55/minetest/issues
22:14 RealBadAngel   there was a pull to remove registered craft recipes
22:15 RealBadAngel   imho very needed one
22:15 sapier         tba this is a feature addon I don't think this should be added after feature freeze
22:15 sapier         -t+r
22:15 kahrl          https://github.com/minetest/minetest/pull/409 Fix glitch ladders -- not sure what this does exactly but sounds non critical
22:16 celeron55_     it's also somewhat controversial
22:17 sapier         kahrl #740 fixes issues where minetest would crash if those cases happened atm
22:17 sapier         #739 fixes a bug I added when adding lua function traces
22:18 kahrl          https://github.com/minetest/minetest/pull/426 this seems to change the liquid update code a lot so probably not suitable for a feature freeze
22:18 sapier         but I didn't test that code still it's small and looks correct
22:18 RealBadAngel   anyways, celeron55_ im goin to start separate game project, whenever you like it or not. im too old for youngster to dictate what i can do. i will publish links for the project and let community decide.
22:19 RealBadAngel   i think it ends the discussion
22:19 kahrl          RealBadAngel, that's what he suggested from the beginning 8)
22:19 RealBadAngel   so done
22:19 kahrl          sapier: I'll test #739-#741 and merge them if they seem ok
22:20 sapier         I'd be sad about you leaving community rba
22:20 kahrl          after I go through these other pulls
22:20 kahrl          sapier, I didn't understand that comment as leaving the community?
22:20 sapier         no?
22:21 sapier         ok so I wont be sad
22:21 sapier         :-)
22:21 kahrl          just building an additional gamemode
22:21 sapier         too many ppl left due to stupid issues the last year
22:22 ShadowNinja    Feature freeze now? How long? /topic?
22:22 kahrl          https://github.com/minetest/minetest/pull/439 needs a rebase
22:22 RealBadAngel   kahrl, but that wasnt the real point, i just showed with this once again that we cannot push anything into minetest_game
22:22 sapier         until 0.4.7 release shadowninja
22:22 kahrl          ShadowNinja: 48 hours
22:22 RealBadAngel   sadly i was right
22:23 RealBadAngel   but shit, there are many folders in the world
22:23 sapier         yes RBA I guess we need improved support for multiple gamemodes
22:23 ShadowNinja    Hmmm, sounds too short...
22:23 kahrl          ShadowNinja: as hmmmm said there were 5 days without any activity
22:24 sapier         there even was a day without comments in minetest-dev ;-)
22:24 kahrl          if a really critical bug is found the duration can be extended
22:25 sapier         any volonteers for maintaining a stable branch adding critical fixes 2 weeks after release?
22:26 kahrl          sapier: it'd be easier to just make another release
22:26 sapier         but usually we don't do that ;-) have a look at 0.4.6 there are some critical bugs in
22:27 kahrl          yeah that was bad
22:27 sapier         but it's only a suggestion
22:28 sapier         e.g. releasing 0.4.7 and about two weeks later 0.4.7.1 containing only bugfixes for critical problems missed
22:28 celeron55_     nobody should have any need to think like sapier now said
22:28 celeron55_     a release should always be made if bad bugs are fixed
22:28 celeron55_     and no, that's fucking horrible
22:29 celeron55_     there are already three version levels, the last one is "patch" level which is the one that can be arbitrarily incremented
22:29 sapier         I totaly agree with you celeron that it should be ... but most time it isn't
22:29 celeron55_     well then make it not be
22:29 celeron55_     i mean, be
22:30 sapier         I don't see how we can do this without having ppl test our versions ... maybe by releasing a 0.4.7 rc1 about a week prior release?
22:30 celeron55_     just release 0.4.8 immediately after if 0.4.7 turns out to be crap
22:30 celeron55_     it's not like we'd run out of integers...
22:30 RealBadAngel   most stable code turns out to be unstable when goes public
22:31 sapier         my experience is that doesn't happen but maybe I'll be prooven wrong in future ;-)
22:31 RealBadAngel   abusers are able to do things simple coders cannot imagine ;)
22:31 celeron55_     sapier: say that to those who released 0.4.6
22:32 sapier         I assumed those ppl are in here :-)
22:36 kahrl          most of the other pull requests are enhancements...
22:36 kahrl          next fix: https://github.com/minetest/minetest/pull/670
22:37 kahrl          "Fix build on big endian architectures" did anyone test this / look at it?
22:38 kahrl          they didn't provide the link celeron asked for :|
22:41 celeron55_     it's stupid because experienced programmers don't care to explain their doings, and noobs don't know what they're doing so they can't explain it, and in the end both look the same
22:42 kahrl          I think a better solution would be to use TestBigEndian.cmake in order to not rely on compiler specific things
22:42 sapier         670 only changes #if to #ifdef
22:44 kahrl          oh and they change irrUString.h which should ideally be left untouched
22:44 kahrl          in case somebody wants to update that file from upstream in the future
22:46 kahrl          hmm apparently TestBigEndian.cmake does not handle cross compilation
22:52 kahrl          putting #670 on hold for now
22:53 kahrl          https://github.com/minetest/minetest/pull/706 HUD Bugfix -- is actually an enhancement
22:56 kahrl          https://github.com/minetest/minetest/pull/730 modified print format codes to use platform specific 64-bit codes (mingw)
22:57 kahrl          should probably be done but are inttypes.h and the PRIu64 etc. macros available on all supported platforms?
22:58 sapier         afaik inttypes.h is standard ... but not sure if all of it's content is standard too
22:59 kahrl          perhaps the places that use %llu should use ostringstream instead of sprintf
23:11 PilzAdam       kahrl, re: fix glitch ladders: last time I mentioned it here people said they want it to stay, although I prefer to fix it
23:12 PilzAdam       I cant remember who "people" are, though
23:12 PilzAdam       IIRC Jordach
23:18 kahrl          it's a controversial thing in any case
23:19 kahrl          not suitably for merge during a feature freeze
23:19 kahrl          suitable*
23:19 PilzAdam       btw: whoever will bump the version, do not forget to update lua-api.txt too
23:20 PilzAdam       and maybe update the link at the top of it too
23:23 kahrl          any objections? https://gist.github.com/kahrl/5688612
23:24 kahrl          this fixes #730 without relying on the existence of certain macros in inttypes.h
23:24 kahrl          which might be fine to rely on but I'm not sure
23:25 kahrl          hold on this is missing a percent
23:26 sapier         std::ios_base::binary? why? (I don't know if it's right or wrong i just don't understand it)
23:26 PilzAdam       I really dont know much about strings in c++, but when I look at this then I dont want to learn it :-)
23:27 kahrl          sapier: it's used almost everywhere that ostringstream is used and doesn't really hurt
23:27 sapier         then this is a wrong conlusion pilzadam strings in c++ are quite powerfull so it's usage is a little bit more difficult than char* ;-)
23:27 kahrl          makes the output of the stringstream more predictable in any case
23:28 Exio           sapier: it is called "i already know brainfuck, i don't want to learn whitespace"
23:28 PilzAdam       sapier, Im used to the java.lang.String class
23:29 sapier         I guess it's somehow similar but I don't use java regularly
23:29 Exio           lol PilzAdam
23:30 kahrl          with the missing percent sign: https://gist.github.com/kahrl/5688630
23:31 PilzAdam       its good as long as it works and you dont change the ouput
23:32 PilzAdam       does this "std::setprecision()" have any effect outside of the string?
23:32 kahrl          only if you append to the same ostringstream afterwards
23:33 kahrl          which doesn't happen (the scope is closed a few lines below)
23:33 kahrl          I changed a tiny thing in the output
23:33 kahrl          "drawtime=" is now "drawtime = "
23:33 kahrl          for symmetry with the other variables
23:42 kahrl          sapier: are you OK with https://github.com/minetest/minetest/pull/742 being merged into upstream?
23:43 kahrl          it will cause rebase conflicts with next_gen_main_menu but they will be easy to fix
23:43 sapier         yes add it I'll fix the conflicts
23:44 kahrl          ok anyone else any objections (also to https://gist.github.com/kahrl/5688630)
23:45 hmmmm          huh?  what's the point of that last one
23:45 Exio           fixing warnings
23:45 hmmmm          ah
23:46 kahrl          hmmmm: mingw's (sn)printf doesn't support %llu
23:46 hmmmm          that's a real shame
23:46 kahrl          needs annoying compatibility macros instead which I'm not sure exist everywhere
23:46 hmmmm          they are standard for POSIX systems
23:46 kahrl          mingw needs %I64u
23:49 kahrl          I'll merge the gist then and then #742 and then I'll look at #739-#741
23:55 Exio           why is #562 won't fix?
23:57 sapier         at least scriptapi separation added some changes to mutexes
23:58 Exio           kk
23:58 sapier         so system wide most likely would break build
23:58 sapier         i didin't even know about that issue until now