Time  Nick           Message
05:02 hmmmm          heh.  i could've been OO-y and made Ore into a base class that has a pure virtual function to create ores based on which type of ore it is, scattered, 3d perlin noise, sheet, or clay, and all this other nonsense
05:03 hmmmm          but i am resisting that idea and leaving it as a simple struct and have the several generateOres methods of Mapgen
06:52 celeron55_     < hmmmm> why do the dungeon generation in lua?
06:53 celeron55_     because i had such code already lying around and didn't want to bother messing too much with the engine
06:53 celeron55_     too much trouble figuring locations for chests and mobs and so on
06:54 celeron55_     and as the player can move so slowly and not in a too large area, the world just generates in the first seconds and then it's done
06:55 celeron55_     so there are a LOT of reasons; feel free to fix the reasons, but currently they are very, very valid
06:55 celeron55_     also, the dungeon layout for this game is very game-specific
06:57 celeron55_     i can see why people here aren't inventing anything new; everybody just directly goes bashing some technical details and don't care about anything else
06:57 celeron55_     it's not rewarding at all to be creative
07:04 hmmmm          hey look; that's fine and all, i was just wondering why you'd do that because it struck me as odd
07:04 celeron55_     i'm trying to believe you, but i think it's way too obvious
07:25 hmmmm          actually speaking of which
07:25 hmmmm          i noticed you used calcLighting() which is pretty wasteful if everything's the same node, setLighting() should be used in this circumstance
07:59 sapier         I've got a problem with cleaning up scriptapi design, I try to remove redundant therefor possible out of sync information in scriptapi ... but those data spread has resulted in some api limitations. Thus cleaning up this mess will result in (minor) api changes ... at least atm I don't see a way to avoid this :-(.
08:01 sapier         examples always help: currently minetest.env:dosomething  ---> after cleanup: minetest.env.dosomething
08:03 sapier         why? atm there is a envref object pushed to lua stack containing environment information and a separate server object containing server information ... envref is a member of server and can be accessed by server thus this information is redundant within lua stack. There never will be more than one environment per server object.
08:04 sapier         any ideas?
08:18 celeron55_     a minor change? lol
08:18 celeron55_     that will break absolutely everything
08:18 sapier         minor change meaning of syntax not of effect
08:19 celeron55_     you're basically modifiying the first parameter of every minetest.env function
08:19 celeron55_     it's huge
08:19 celeron55_     (removing the first parameter)
08:20 sapier         yes ... result of decision to unlink environment from server within lua stack
08:20 celeron55_     you will need to fully preserve compatibility with the existing api; there are no excuses about that
08:20 celeron55_     and it needs to be clean because it's going to stay
08:21 celeron55_     if you can do the addition of the self parameter version of every function without additional code for every function, then it' fine
08:21 celeron55_     +s
08:21 sapier         atm I see no way to clean up the lua integration mess while preserving 100% compatibility that's why I'm asking  for ideas
08:21 celeron55_     otherwise it will be insane
08:22 celeron55_     i think you will be able to do it
08:22 celeron55_     but not sure exactly how; some metatable trick possibly
08:22 sapier         maybe I'll be able to add a workaround just ignoring first parameter this keeping compatibility
08:23 sapier         there's no neet for environment to be a parameter at all
08:23 sapier         I don't know why this was done that way
08:23 celeron55_     i know there is no need, but you can't break the API
08:23 celeron55_     the only reason to have it now is compatibility, and that is a strong reason
08:24 sapier         I know too ;-) preserving this insane decision will be lots of additional work ;-)
08:24 celeron55_     more like zero work
08:24 sapier         If this is done can we mark minetest.env: syntax as deprecated to be able to remove this hmm lets say with minetest 0.9 ;-)
08:26 celeron55_     of course
08:27 sapier         cleaning up scriptapi is a pain ... it involves changes throughout minetest ... most time minor ... but often difficult to understand how this may have grown
08:27 celeron55_     altough why do you want to keep the .env?
08:27 celeron55_     they all could just be directly in minetest
08:28 sapier         of course woult be even more easy to implement
08:28 sapier         at least more clean
08:28 celeron55_     it will be easier because if a function is called in env, you can always assume it has the self parameter and just throw that away and call the same function in minetest without it
08:28 celeron55_     env could be implemented directly in builtin as lua as a wrapper like that
08:29 sapier         env isn't even needed
08:29 sapier         my current aproach splits scriptapi in a top and bottom half
08:29 celeron55_     get rid of it then; no need to be subtle when you're ripping everything apart anyway
08:29 sapier         top half containing no language reference at all and reducing header linkage to a very small extent
08:30 sapier         bottom half containing everythin only usefull for lua
08:30 sapier         ok so bee it remove .env and add a .env: workaround
08:30 celeron55_     what are these halves you are talking about? halves of a file?
08:30 celeron55_     halves of documentation?
08:31 sapier         both ... top will be only relevant for core developers
08:31 sapier         while bottom defines interface towards lua
08:31 sapier         e.g. the env thing
08:31 celeron55_     oh so one is the _raw stuff and that kind of thing, and other is that which is directly in the mod-facing API?
08:32 sapier         no not exactly top is where all those fct's are that were called scriptapi_
08:32 sapier         bottom everything with l_
08:33 celeron55_     ah that way... i woulnd't call that top and bottom; it's the internal and external interface or something
08:33 sapier         scriptapi_ is only used from within core (at least almost ... there's always one place where this isn't true :-( )
08:34 sapier         you're right this way of calling it is better
08:34 sapier         with this design change it might even be possible to replace lua by vb (at least if someone was willing to invest that time ;-))
08:35 sapier         an of course if there is any change in external interface core doesn't need to be rebuilt
08:36 sapier         let me be more precise core objects only of course
09:19 celeron55_     hmmmm: regarding to changing mapgen, i think we could use such a policy that changing anything else except the sufrace shape is OK
09:20 celeron55_     and changing surface content is strongly discouraged (like modifying biome noise parameters or so)
09:20 celeron55_     but not entirely disallowed
09:22 celeron55_     that is regarding to V6; other policies can be taken into use for other mapgens when needed, but the default mapgen should use this so maps don't spontaneously get odd square cliffs
09:23 celeron55_     it's very limiting but i guess people expect it
11:38 rark           Hi. Is the bug about nodes being overwritten if the undersurface of a pointed thing is builable_to a known bug?
11:39 PilzAdam       its wanted behaviour?
11:39 rark           No, it's a problem.
11:39 rark           You see, if this occurs on a chunk boundary, in some cases the undersurface is seen as a CONTENT_IGNORE when there was actually something useful there.
11:40 rark           This is especially dangerous with chests being the undersurface, and has already been exploited once or twice on one of the servers I moderate.4
11:40 PilzAdam       ah, that bug
11:40 rark           It also seems like Lua doesn't get called during 100% of placenodes when it should.
11:40 rark           So it's known?
11:41 PilzAdam       the problem there is that Lua gets ignore when there is actually a node
11:42 rark           Yeah. And this ignore is exposed to Lua as oldnode in minetest.register_on_placenode?
11:42 rark           It also seems that during testing, sometimes the register_on_placenode callback isn't called at all during node placement. This happens, maybe, 0.1% of the time...
11:43 PilzAdam       that may be server lag
11:43 rark           No, since the node itself does get placed.
11:43 PilzAdam       are you sure that it isnt only node placement prediction of the client?
11:44 PilzAdam       since the whole node placement is done in Lua
11:44 rark           Yeah, since it appears to all other players as such.
11:44 rark           Also, it looks like the lines in server.cpp regarding unpacking the pointed thing somehow do get called.
11:47 rark           scriptapi_item.cpp: lines 255-256 look like they might be the cause of the failure. I'll do more testing.
12:57 rarkenin_      afk
13:51 proller_       https://github.com/minetest/minetest/pull/563
14:05 proller_       maybe make finite liquids enabled by default?
14:05 PilzAdam       no
14:06 proller_       ok. i make checkbox for it
14:06 proller_       in pull ^^
14:15 proller_       ready!
14:21 proller_       and.. maybe make servers without password in list upper than protected?
14:41 celeron55_     proller: by the way, make sure to read this discussion here http://irc.minetest.ru/minetest/2013-03-23#i_2949424
14:45 proller        im ready to use it
14:47 celeron55_     also, i gave you push access to upstream; just make sure at least one other person with such access agrees to whatever you commit before doing so
14:48 celeron55_     (for extra clarity, the list is... umm... here: https://github.com/minetest?tab=members)
14:50 celeron55_     #563 seems good, and you're free to mess around with the indev mapgen as long as you don't flood the commit history with that
14:51 proller        wow, thank you
14:52 celeron55_     (maybe squash and push changes only a few times a week or so; quite useless to update it all the time)
14:53 proller        ok
14:55 proller        about https://github.com/minetest/minetest/pull/563 - everyone agree with default multiplayer tab with online servers list ?
14:56 celeron55_     the commit message should be put on single line and started with a capital letter 8)
14:56 celeron55_     (because it fits on a single line)
14:56 proller        ok
14:57 rarkenin_      Maybe if an autogenerated name is used and the server gives an access denied, a new number should be generated?
14:57 celeron55_     (change it to something like "Random guest name and finite liquid checkbox"
14:57 celeron55_     )
14:58 celeron55_     the default tab should probably be voted on on the forums
14:59 celeron55_     personally i don't really care
14:59 celeron55_     the user sees it only one time per installation
14:59 proller        rarkenin_, its harder, needs var to store, and now - very low chance to get access denied
14:59 thexyz         um
15:00 thexyz         can't we replace https://github.com/minetest/minetest/pull/563/files#L1R94 with serverlist_show_available(USE_CURL) ?
15:00 rarkenin_      OK.
15:00 rarkenin_      Maybe 5 or six digits, then? (Just my paranoia and skewed statistics and probability analysis)
15:01 celeron55_     4 is plenty enough
15:01 proller        rarkenin_, now only ~100 active players with names
15:02 celeron55_     i do think some people will not like the change, but i can't really guess how many (if they often intend to input a name, but forget, and then realize they're guest3244)
15:03 celeron55_     the finite liquid checkbox is also arguable - it really should be per-world
15:03 celeron55_     but so should be the creative and damage ones, but we lack the means to do that
15:04 celeron55_     (that is a thing to be added)
15:09 celeron55_     so if you care, you could design that first and then make the perfect solution
15:10 rarkenin_      About the forgetting a name part, the last name used on any server should be already in the textbox.
15:10 proller        celeron55_,  maybe later, now i want to finish float lands
15:10 celeron55_     rarkenin_: yes it should... so that applies only to people forgetting to input a name who regularly wipe minetest's config 8) that should be very rare
15:11 proller        and try many weather things
15:13 proller        thexyz, USE_CURL can be empty string
15:14 proller        maybe not with current cmake
15:23 proller        563 merged, my first commit 8-)
15:27 proller        ups. using github merge - bad way ?
15:27 thexyz         who cares
15:27 proller        it make 2 commits
15:29 PilzAdam       I care
15:32 celeron55_     PilzAdam seems to be the only one who cares these days :P
15:33 celeron55_     maybe just stop caring, it's much easier that way
16:14 celeron55_     hmm
16:14 celeron55_     we should probably move more stuff to common as this thread suggests: http://forum.minetest.net/viewtopic.php?id=5210
16:15 celeron55_     dunno
16:15 PilzAdam       yea
16:15 PilzAdam       basically everything that is in minetest_game can be moved to common
16:15 celeron55_     well not everything; there is no need to move legacy in there, for example
16:15 celeron55_     except for reasons that are not valid and should be solved by other means
16:17 PilzAdam       maybe we can move everything except legacy to common, and start the survival and build games
16:18 PilzAdam       and just leave minetest_game as is (except the updates in common)
16:18 celeron55_     i think everything but give_initial_stuff and legacy could be moved (the first one because it's so static and unconfigurable and simple - any game can do that with no effort)
16:19 celeron55_     and then handy aliases for items should be added to default, so that there is no need for legacy for them
16:20 PilzAdam       is give_intial_stuff actually used by anyone?
16:20 celeron55_     dunno; it's useful for quick testing of non-creative mode
16:20 celeron55_     (it's the reason i added it, i think)
16:22 PilzAdam       can you create the survival and build repos so I can start doing something?
16:28 celeron55_     hmm, is "survive" worse than "survival"?
16:28 PilzAdam       yes
16:28 celeron55_     because survive is a direct verb like build, and i'd imagine naming other games would be easier in this way than as trying to make them non-verbs
16:29 PilzAdam       survival is a well known term in games
16:29 PilzAdam       everything else can be a verb
16:31 celeron55_     now they exist
16:33 PilzAdam       can you add TeamMinetest to them?
16:33 celeron55_     ... 8)
16:33 celeron55_     done
16:33 PilzAdam       starred :-)
16:39 PilzAdam       game.conf and readme added
16:40 PilzAdam       now move the mods from minetest_game to common
16:41 khonkhortisan  rubenwardy found the cloning device
16:49 PilzAdam       okay done
16:50 PilzAdam       okay, now lets give them their renewable lava back in the build game
16:56 PilzAdam       celeron55, there is a problem with dependencies in common mods
16:57 PilzAdam       they are simply ignored
16:57 PilzAdam       with common_mods = default, creative the hand gets overriden correctl by creative; with common_mods = creative, default the default definition of the hand is used
16:58 PilzAdam       and also the common mods are marked as missing in the GUI
16:59 sapier         I'm not quite used to mod loading code is this as mixed througout core as mod api is?
17:05 celeron55      PilzAdam: yes, they need to be hand sorted
17:06 PilzAdam       can this be fixed somehow?
17:07 celeron55      probably, but i'm not too eager to mess around with doserj's code
17:07 celeron55      it isn't exactly trivial
17:07 celeron55      just put them in the right order for now 8)
17:08 sapier         is there any reason why auth handling is implemented in lua not within core?
17:08 celeron55      because handling it in lua allows easily implementing other authentication backends by replacing the handler
17:09 celeron55      in case you haven't noticed, that is fully supported
17:09 sapier         ok sounds reasonable
17:10 sapier         I was just wondering because this creates some strange code for transforming auth data lua <-> c
17:10 celeron55      http://www.minetest.net/extend.php <- i need a good and short guide for installing mods and texture packs for linking to from this page
17:12 celeron55      the wiki.minetest.com stuff is kind of unprofessional
17:12 PilzAdam       also write about games there
17:12 celeron55      i guess these are currently the best ones http://minetest.net/wiki/doku.php?id=installing_mods http://forum.minetest.net/viewtopic.php?id=1592
17:12 rubenwardy     celeron55: http://multa.bugs3.com/minetest/manual/game/custom.html
17:13 PilzAdam       window!?
17:13 PilzAdam       +s
17:13 rubenwardy     yeah :(
17:13 rubenwardy     It is not as good as the http://minetest.net/wiki/doku.php?id=installing_mods one
17:13 celeron55      those screenshots are incredibly hard to understand
17:14 thexyz         MineTest =(
17:16 khonkhortisan  Why are the other mods blurred out? And why can I read them anyway?
17:35 hmmmm          what's with the uptick in celeron activity
17:35 hmmmm          thought you hated minetest
17:36 hmmmm          i agree that if it's better with no speed regressions, then great, we'll add it, but he wants to shitify things up with too much OO
17:36 celeron55      you're free to propose better organization for the code once the code pops up
17:37 hmmmm          also, the noise functions were a bottleneck, but not the only bottleneck
17:38 hmmmm          it just so happens that the lighting change sped things up so much because of the previous noise work.  if the noise bottleneck is gone, and the lighting bottleneck is still there, a similar absolute speedup in lighting will make a much larger relative difference
17:38 hmmmm          that's just how math works
17:38 hmmmm          if i did it in the opposite order people would be jizzing themselves over the improved noise instead of the improved lighting
17:40 sapier         does anyone have objections against moving all lua <-> c++ functions to one file?
17:40 hmmmm          if it seems like i'm so oposed to prestidigitator, that's because he's basically undoing everything i just did
17:40 sapier         type conversion
17:41 sapier         e.g. readnode/pushnode or read_v3f/push_v3f
17:42 hmmmm          dunno, doesn't really matter, it's just an additional organization change
17:42 hmmmm          i'm not sure if it'll be as big of an improvement as splitting the files was
17:42 sapier         it's relevant in order to reduce number of rebuilt files and of course to cleanup scripting api
17:42 hmmmm          sure, why not
17:43 sapier         believe me it is as those functions create depentencys between almost all mods ... moving conversions to one file will at least remove cross dependencys
17:43 sapier         mods meaning inventory/entity/object ...
18:30 RealBadAngel   hi
18:30 RealBadAngel   https://github.com/minetest/minetest/pull/564
18:30 RealBadAngel   6d facedir is ready
18:54 PilzAdam       maybe we should change the name of minetest_game in minetest.conf, so people dont think thats it is the "official" game or so
18:55 PilzAdam       *game.conf
18:55 VanessaE       change it to what?
18:55 PilzAdam       minetest_game?
18:55 VanessaE       oh
18:57 PilzAdam       other suggestions?
18:57 VanessaE       none at present.
19:04 celeron55      "Traditional Minetest 0.4"
19:05 celeron55      8)
19:05 VanessaE       I think if you start giving things new names, it may be about time to move away from calling it "minetest".
19:05 VanessaE       (except the engine)
19:06 PilzAdam       why does the name of the folder and not the name of game.conf appear behind the worlds (in the [])?
19:06 celeron55      it's called Minetest - it's just referring to the whole, consisting of multiple "micro-games"
19:06 celeron55      PilzAdam: maybe that could be changed
19:07 hmmmm          if proller was added to the committers, then why not RBA?
19:08 VanessaE       I could agree with RBA being on the team.
19:09 hmmmm          he does a lot, i'd have to say enough to be considered a core developer
19:09 hmmmm          ugh why can i not focus on anything
19:09 hmmmm          my mother is going on with her squeaky voice in the other room about i have no idea what, my back hurts, and i can't even frigging type some code out
19:10 hmmmm          i need to start drinking
19:10 VanessaE       haha
19:11 celeron55      or sleeping, and buy some kind of ear protectors
19:13 celeron55      i guess i'll add RBA too... RealBadAngel: you probably know the rules of committing to upstream
19:17 celeron55      (in the most compacted form: agreement of somebody else on the team, and a good-looking commit message)
19:18 hmmmm          and don't change upstream history if it's over 10 minutes and you're sure nobody else just pulled
19:18 PilzAdam       and no merge commits :-)
19:18 hmmmm          also agree with that
19:22 celeron55      i hate github for not providing a button to rebase something instead of merging
19:22 hmmmm          ugh, every time i have to write it out, i get this urge to make a typedef "MMVM" for ManualMapVoxelManipulator
19:22 celeron55      i guess they want to grow their social crap by marking merges with those
19:24 celeron55      i once hated merge commits, but now i hate avoiding those
19:24 PilzAdam       if you merge with merge commits the actual commit is placed somewere in the history
19:25 celeron55      it's placed to the point of time in history where it was coded, yes
19:25 PilzAdam       so the whole history is messed up
19:25 hmmmm          yes, that's bad, we don't want that
19:29 RealBadAngel   sorry was away to do some shopping
19:30 RealBadAngel   i think i know the rules
19:31 RealBadAngel   my latest commits are propably lookin better than first ones ;)
19:32 RealBadAngel   now im going to make aeroplane model
19:33 RealBadAngel   that will be used as test object for multitextured nodeboxes
19:34 troller        hmmmm, now too hard add new noiseparams, 5-6 places to write
19:35 hmmmm          jeez, when i came up with the noiseparams, i thought that was a set that everybody could agree on
19:36 hmmmm          noiseparams is supposed to just be that - the parameters that would usually be used in a perlin noise call - nothing else
19:36 hmmmm          i think people are too broadly interpreting what a noiseparams is, they think that it's just every aspect that describes noise
19:37 celeron55      http://dev.minetest.net/Git_Guidelines
19:38 troller        nparams_v6_def - seems not used, they gets from config
19:38 PilzAdam       you actually can merge --ff-only
19:38 celeron55      well, yeah, but in the general case that will usually fail
19:39 hmmmm          i do a cherry-pick against a current branch
19:39 celeron55      setting up remotes for all random contributors is a PITA
19:39 hmmmm          i don't have to..
19:40 PilzAdam       https://gist.github.com/PilzAdam/5229094
19:40 hmmmm          git checkout -b some_shmucks_branch && git fetch http://the_shmucks_repo.git some_shmucks_branch && git checkout master && git cherry-pick the_shmucks_commit
19:41 hmmmm          and troller, nparams_v6_def_* are used to initialize the associated MapgenParams structure
19:41 hmmmm          because what if they're not modified by any config at all
19:42 celeron55      compared to "wget https://github.com/minetest/minetest/pull/564.patch | git am -", that's quite worky
19:42 celeron55      + -O-
19:42 celeron55      ehm... -qO-
19:42 PilzAdam       if you have the contributors list once setted up, you can always fetch the repo without copying the URL all the time
19:42 hmmmm          the understood rule of noiseparam defaults is, if you add something new, you have a default setting in the config, and a mirrored version of the struct with an nparams_<mgnamehere>_def_* in your file
19:49 hmmmm          hey sapier are you around?
19:49 RealBadAngel   so, do you folks agree to merge https://github.com/minetest/minetest/pull/564 ?
19:49 hmmmm          NO!
19:49 hmmmm          absolutely horrible
19:50 RealBadAngel   :) too much rotation freedom?
19:51 hmmmm          ha ha, i like your axisdir = facedir >> 2; facedir &= 0x03; that's like something i would write and everybody else would hate
19:52 RealBadAngel   that was logical solution
19:52 RealBadAngel   this way change doesnt break any older stuff
19:52 hmmmm          indeed
19:53 hmmmm          sapier, nevermind, i figured it out.. i was just going to ask where to put EnumString definitions in scriptapi
19:53 RealBadAngel   it just extends it
19:55 RealBadAngel   btw i think we need some bitwise library for Lua
19:55 troller        hmmmm, why need to mirror params?
19:56 RealBadAngel   http://bitop.luajit.org/
19:56 RealBadAngel   like this one
19:56 hmmmm          troller, to initialize MapgenParam structures
19:56 VanessaE       RealBadAngel: luajit has that.
19:56 hmmmm          yeah but now we need to make lua have that
19:56 hmmmm          i agree that we need bitop, but we need to add the extension to our own copy of lua as well
19:56 hmmmm          it's either that, or we remove bitop from luajit
19:56 RealBadAngel   its optional library for luajit
19:56 hmmmm          doesn't matter what way it happens, what's important is that it's consistent
19:57 hmmmm          if something can work in luajit but not work in lua, then it's unacceptable
19:57 RealBadAngel   when we move to luajit why care about old lua?
19:57 hmmmm          because we never moved to luajit
19:57 hmmmm          we're just using it if we're lucky
19:57 thexyz         why not move to it then?
19:57 hmmmm          because not everybody has luajit
19:57 thexyz         we can bundle it
19:57 RealBadAngel   not yet, but move will be not an option but switch to
19:58 hmmmm          luajit isn't available for all architectures
19:58 RealBadAngel   i mean together with move to jit add also bitop
19:58 hmmmm          it's not universal like lua is
19:58 hmmmm          yes, i'll agree with that
19:58 hmmmm          we add bitop to the bundled lua, that'll be okay
19:58 VanessaE       hmmmm: are you sure about that?  (all archs; it's available on everything MT runs on now)
19:59 hmmmm          but imagine, somebody doesn't have luajit, "whoops, CrazyMod 4.0 doesn't work, screw you"
19:59 VanessaE       heh
19:59 hmmmm          it's not hard to add the extension
19:59 VanessaE       your point is valid
19:59 hmmmm          just do it and please don't break compatibility
20:00 PilzAdam       there are no luajit win32 builds :-(
20:00 thexyz         x86, x86_64, arm, ppc, mips
20:00 hmmmm          we never officially added it
20:00 RealBadAngel   hell it works even on PS3 and xbox
20:00 troller        how to show players ip? on server, want to ip ban
20:00 thexyz         sure, minetest will run on all of them
20:00 thexyz         and on some other which are not supported by luajit
20:01 hmmmm          it doesn't run on SPARC, PA-RISC, and POWER
20:01 VanessaE       troller: just ban, it works by IP anyway
20:01 VanessaE       Power?
20:01 hmmmm          it's not 100% compatible with PPC if that's what you're saying
20:02 VanessaE       hmmmm: the overview says jit works on PPC
20:02 thexyz         um, sure, sparc
20:02 thexyz         who uses anything other than x86/arm those days?
20:02 VanessaE       thexyz: some supercomputing facilities? ;)
20:02 hmmmm          that doesn't matter, the point is, why limit ourselves when we could've just added the extension to lua?
20:03 thexyz         because it costs time, and no one will use it?
20:03 hmmmm          if you notice, my platform-specific thread code supports everything except for IRIX and OpenVMS
20:03 thexyz         that's great
20:08 VanessaE       I personally see no reason to keep interpreted Lua at all - just bundle JIT and throw away the interpreter version
20:08 VanessaE       let the game link against a system-wide interpreter if the user wants to
20:08 VanessaE       s/game/engine/
20:08 thexyz         well, the reason is clear
20:08 thexyz         it's not supported at some platforms
20:09 RealBadAngel   some very exotic ones
20:09 VanessaE       it's not supported on like 0.1% of platforms the game can run on
20:10 RealBadAngel   so maybe keep old one just in case?
20:10 RealBadAngel   propably bitop would work with old lua too
20:10 VanessaE       Linux, BSD, OSX, Windows; AMD64, x86, PPC
20:10 VanessaE       this covers 99.9% of all PCs
20:10 PilzAdam       there are no official builds for windows
20:10 VanessaE       so? bundle and compile it.
20:10 hmmmm          i know we support the big four, but my personal standard is to support everything possible
20:10 thexyz         PilzAdam: so..?
20:10 RealBadAngel   even my ZX Speccy? ;)
20:10 thexyz         what stops us from building it?
20:11 VanessaE       hmmmm: that's a laudable goal, but is it realistic?  I'm just trying to avoid project bloat.
20:11 PilzAdam       I have to build it when cross compiling wich takes time...
20:11 hmmmm          if LuaJIT is dynamically linked, then there would be no additional bloat...
20:11 RealBadAngel   offtopic: can i merge 6d facedir? :)
20:11 thexyz         well, we can drop all additional libraries which are incompatible with pure lua
20:11 hmmmm          yes
20:12 VanessaE       hmmmm: but then PilzAdam's "there's no win build" argument comes into focus.
20:12 thexyz         and still gain perfomance increase
20:12 * RealBadAngel is lookin for the button :)
20:12 PilzAdam       it isnt really an argument
20:12 hmmmm          noooo
20:12 hmmmm          realbadangel, don't
20:12 thexyz         RealBadAngel: i'm pretty sure this button is called git cherry-pick
20:12 hmmmm          rebase it and push
20:12 hmmmm          don't press the button
20:12 VanessaE       RealBadAngel: NOT THAT BUTTON!  NOT THE RED ONE!
20:12 RealBadAngel   hehehe
20:13 hmmmm          not the green one
20:13 PilzAdam       *GREEN
20:13 VanessaE       (or green in this case)
20:13 hmmmm          press the red one
20:13 VanessaE       way to kill a joke, guys :)
20:13 thexyz         VanessaE: what's that "no win build" about?
20:13 VanessaE       thexyz: well it you dynamically link, you're implying a shared, third-party-supplied library, right?
20:13 VanessaE       just as you need OpenAL to get sound
20:14 thexyz         you can package them with the main binary
20:14 thexyz         like, .dll-s
20:14 PilzAdam       RealBadAngel, why the merge commit?
20:14 VanessaE       thexyz: that's what I was saying should be done, but that's not the impression one gets
20:15 PilzAdam       RealBadAngel, you broke rule 4: http://dev.minetest.net/Git_Guidelines
20:15 celeron55      this is quite much the reason why i haven't given RBA upstream access; he tends to do first and think second
20:15 celeron55      hopefully that'll improve
20:15 thexyz         not like anybody cares though
20:15 RealBadAngel   damn :(
20:15 thexyz         umm, *cares*
20:15 thexyz         just don't do that again
20:15 PilzAdam       git rebase --hard HEAD~1
20:16 hmmmm          10 second rule
20:16 PilzAdam       *minutes
20:16 hmmmm          you can still eat^H^H^H rewrite it
20:16 PilzAdam       has anyonre pulled yet?
20:16 hmmmm          of course not
20:16 RealBadAngel   so what i should do now?
20:17 PilzAdam       git rebase --hard HEAD~1; git push -f upstream master
20:17 thexyz         PilzAdam: I've just downloaded zipball to build windows binary :P
20:17 PilzAdam       *reset
20:17 PilzAdam       thexyz, I bet not the official one
20:17 thexyz         guess it doesn't count
20:17 thexyz         as it's zipball
20:17 hmmmm          i'd say part of it is the MinetestBot in #minetest, people think "OOOOOOOH NEW COMMIT, I NEED TO UPDATE"
20:17 thexyz         it knows nothing about history
20:17 thexyz         we should make it wait 10 minutes before announcing new commit
20:17 thexyz         sfan5: ^
20:18 sfan5          who cares about that?
20:18 sfan5          thexyz: ^
20:18 thexyz         sfan5: see discussion above
20:18 RealBadAngel   git rebase --hard HEAD~1
20:18 PilzAdam       people who have to pull a force-updated repo
20:18 thexyz         ^^^^^^^
20:18 PilzAdam       RealBadAngel, *reset
20:18 sfan5          *reading*
20:18 RealBadAngel   are you sure about it?
20:18 VanessaE       if you track git, you should be prepared for the occasional oddity.
20:19 VanessaE       why do we care about merge commits?
20:19 VanessaE       I mean seriously?
20:19 hmmmm          read up
20:19 hmmmm          we explain why in great detail
20:19 PilzAdam       because the actual commits are placed somewere else in the history
20:19 VanessaE       um
20:20 VanessaE       then github is doing something weird, because I see them in the order they're committed
20:20 PilzAdam       wich goes against the whole idea of a history
20:20 thexyz         meh, i hope we won't end with empty repo
20:21 RealBadAngel   PilzAdam, git rebase --hard HEAD~1  this doesnt look like a valid command
20:21 PilzAdam       it is
20:21 PilzAdam       well, reset instead of rebase
20:21 RealBadAngel   realbadangel@Haven:~/Git/minetest$ git rebase --hard HEAD~1
20:21 RealBadAngel   Usage: git rebase [--interactive | -i] [-v] [--force-rebase | -f] [--no-ff] [--onto <newbase>] (<upstream>|--root) [<branch>] [--quiet | -q]
20:21 PilzAdam       git reset --hard HEAD~1
20:22 PilzAdam       1 minute left
20:22 RealBadAngel   HEAD is now at 2318d19 Merge pull request #563 from proller/guest
20:22 PilzAdam       git cherr-pick b1c2bed8b3b42f8cd709b9343ffc82c5179197eb
20:22 PilzAdam       git push -f origin master
20:23 PilzAdam       *cherry-pick
20:23 RealBadAngel   done
20:24 RealBadAngel   now what?
20:24 PilzAdam       push
20:24 VanessaE       push -f.
20:24 PilzAdam       like normal push but with -f or --force param
20:24 thexyz         or with +
20:25 thexyz         git push origin +master
20:25 PilzAdam       ummmm....
20:25 PilzAdam       why have you opened a pull request?
20:26 RealBadAngel   closed it
20:26 RealBadAngel   i thought i should pull it from github page
20:27 PilzAdam       should I reset it for you?
20:27 RealBadAngel   i dont have here cloned original repo but my fork
20:27 troller        hmmmm, why not https://github.com/proller/minetest/commit/84e3d807ec4a3a0c9c0a63e3094afcca2d33d52f ?
20:28 thexyz         hm, tried to build in clean environment, msvc2010, still get link error
20:28 thexyz         mapgen_indev.obj : error LNK2001: unresolved external symbol "public: virtual void __thiscall MapgenV6::makeChunk(class BlockMakeData *)" (?makeChunk@MapgenV6@@UAEXPAVBlockMakeData@@@Z)
20:28 hmmmm          troller, that's the first i've seen that
20:29 hmmmm          also thexyz, i figured it out!
20:29 hmmmm          it's because i forward declare BlockMakeData as a class when it's actually a struct
20:29 troller        hmmmm, its try to reduce noiseparams places
20:29 hmmmm          (thanks indivisibleatom)
20:29 thexyz         wow, great
20:29 hmmmm          thexyz, you need to merge this https://github.com/minetest/minetest/pull/561
20:29 RealBadAngel   PilzAdam, if you can help, sure
20:30 RealBadAngel   for next commit i will try to learn some more
20:30 thexyz         hmmmm: can't we just merge it into master?
20:30 hmmmm          ??
20:30 hmmmm          the build fix?
20:30 hmmmm          yeah that's the point
20:31 thexyz         oh
20:31 PilzAdam       done
20:31 hmmmm          erm troller, that doesn't really help matters.. NoiseIndevParams will be partially filled with junk if it is not written to by the config
20:31 hmmmm          actually s/noise/zero/
20:31 PilzAdam       now upstream is pullable again
20:32 troller        hmmmm, but when it can be filled not from defaultsettings.cpp ?
20:33 hmmmm          when someone creates a MapgenParams structure somewhere else in the code
20:36 RealBadAngel   ok, im fixing now screwdriver mod
20:36 PilzAdam       in the current implementation its not possible for stairs and slabs to use the 6d facedir
20:36 hmmmm          wait !?!
20:37 hmmmm          i thought that was tested already to work
20:37 PilzAdam       we need to copy whole item_place_node into the mod
20:37 RealBadAngel   PilzAdam, everything that uses facedir is working with 6d
20:38 PilzAdam       yea, its what needs to be done to make stairs and slabs using it
20:39 PilzAdam       it currently creates a fakestack with upside down node and calls item_place with it
20:39 VanessaE       latest git w/6d + luajit linked in is now deployed to my server.
20:39 PilzAdam       its not possible to pass param2 to item_place
20:40 hmmmm          uh oh
20:40 hmmmm          vanessa, don't be in such a hurry...
20:40 PilzAdam       but if I copy a slightly modified version of item_place_node to I could also fix the slab to full block transformation in creative mod bug
20:40 VanessaE       hmmmm: my server's kinda meant for testing stuff anyway.  if it breaks, I'm not too worried, but everything seems to be normal.
20:41 hmmmm          also, unintentional, but neat.. anyone like this type of ore?  http://ompldr.org/vaHYwcQ
20:41 VanessaE       hmmmm: YES!
20:41 RealBadAngel   PilzAdam, 6d change doesnt interefere with placing nodes code or whatever
20:41 PilzAdam       RealBadAngel, yes
20:42 RealBadAngel   if you want to use the feature just do so
20:42 PilzAdam       Im just talking about how to make slabs and stairs use it
20:42 RealBadAngel   screwdriver :)
20:42 PilzAdam       hmmmm, NO!
20:42 RealBadAngel   also new placement code is needed
20:43 PilzAdam       in what way?
20:45 RealBadAngel   look dir to facedir need changes
20:47 RealBadAngel   old version does care only for y+ axisdir
20:48 RealBadAngel   placement code need to be changed to place up/down pistons for example
20:48 VanessaE       there, refreshed my latest build - something was odd with the commit IDs
20:49 VanessaE       server is now on the very latest git of all four components.
20:49 RealBadAngel   how is the memory usage now?
20:49 VanessaE       let me check
20:49 VanessaE       hold.
20:50 Exio           well, most of devs, what about my "idea" of adding common, at least as git submodule?
20:50 VanessaE       preload item visuals....ON.  512px enabled.  loading.
20:50 Exio           afaik you don't download it by default when just doing "git clone ..."
20:51 VanessaE       client at 6GB resident and counting.
20:52 VanessaE       6.2GB
20:52 VanessaE       loaded.
20:52 Exio           preload item visuals should be called "Ram Killer"
20:52 VanessaE       it'll add a few hundred megs once there's plenty of map loaded around me
20:52 VanessaE       wow
20:52 VanessaE       30 fps
20:52 VanessaE       oh how I missed this
20:53 Exio           hehe
20:53 VanessaE       my memory usage went down by about 500 MB I would guess.
20:53 Exio           try to open minecraf with a 512x texture pack
20:53 Exio           the counter will start using FPM than FPS ^^
20:54 Exio           instead of*
20:54 RealBadAngel   http://i.imgur.com/E51fjSG.png
20:54 VanessaE       I can still outrun the map generator though, but only when I'm in an area with lots of big trees
20:54 RealBadAngel   ^^regarding stairs
20:55 Exio           i can't with default in my bulldozer 6100
20:55 VanessaE       s/map generator/map loader-sender/
20:55 Exio           before, i was able to reach "non generated chunks"
20:55 Exio           now i can barely with super-speed ^^
20:55 RealBadAngel   VanessaE, so 6dfacedir speeded up the game for you?
20:55 VanessaE       RealBadAngel: immensely.
20:55 VanessaE       well, let's call it "significantly"
20:56 RealBadAngel   good :)
20:57 RealBadAngel   i think i will try to connect to your server with hdx
20:57 RealBadAngel   last time my 4gbs wasnt enough to do so
20:57 VanessaE       I'm at 7.3GB used now
20:57 VanessaE       (but a lot of that is map data)
20:58 Exio           is there any way in the config to define "the render distance" of the "limited view distance"??
20:58 PilzAdam       RealBadAngel, when I have the facedir that I get with minetest.dir_to_facedir(), what do I have to do with param2 to get it upside down?
20:58 VanessaE       Exio: viewing_range_nodes_max = 100
20:58 VanessaE       viewing_range_nodes_min = 100
20:59 Exio           thanks ^^
20:59 VanessaE       (that's my default setting, since I don't like the autotuner)
20:59 RealBadAngel   0-3 are old values, means top of the node is facing y+
20:59 RealBadAngel   axisdir is 0 then
20:59 RealBadAngel   you need axisdir=5 to get top facing y-
20:59 PilzAdam       so just add 4?
20:59 RealBadAngel   5*4
21:00 RealBadAngel   range 20-23 will be rotations with top facing y-
21:01 PilzAdam       doesnt work with facedir = 1 and 3
21:01 RealBadAngel   are you using facedirtools?
21:01 PilzAdam       nope
21:01 RealBadAngel   try
21:01 PilzAdam       why?
21:01 RealBadAngel   the tool will show you how to use new facedir
21:02 PilzAdam       I want stairs to be flipped upside down
21:02 PilzAdam       so I add 20 to param2
21:02 RealBadAngel   so set facedir=20
21:02 PilzAdam       it works when the old facedir is 0 or 2
21:02 PilzAdam       but not for 1 and 3
21:03 PilzAdam       if its 1 it acts like it was 3
21:03 PilzAdam       and the other way round
21:04 RealBadAngel   nope, tested
21:04 RealBadAngel   hold on for screenshot
21:05 thexyz         um.. add 20?
21:06 thexyz         lua api doesn't specify how facedir is calculated, so it's unsafe to rely on magic like that
21:06 RealBadAngel   http://i.imgur.com/1ReegtT.png
21:06 celeron55      umm i think i know what PilzAdam is seeing
21:06 thexyz         "implementation-dependent"
21:06 PilzAdam       what I do is minetest.dir_to_facedir(dir) + 20
21:06 RealBadAngel   from left to right
21:06 RealBadAngel   20 to 23
21:07 celeron55      PilzAdam: the stairs rotate to a different direction when looking from the same static position, when they are upside down
21:07 PilzAdam       now place 0 to 3 in front of them
21:07 celeron55      because they rotate with respect to the axis that defines whether they are upside down or upside up
21:09 RealBadAngel   http://i.imgur.com/JQdzx2j.jpg
21:10 RealBadAngel   first row is axisdir y+, second row is y-
21:10 VanessaE       everything is defined relative to where the node's natural "top" face is pointed.
21:10 VanessaE       so in the first row, the natural top face points toward Y+, while in the back row it points to Y-
21:11 celeron55      RealBadAngel: so exactly; looking from up, the other rotates clockwise, the other counter-clockwise
21:11 RealBadAngel   i suggest using facedirdools
21:11 RealBadAngel   strongly
21:12 RealBadAngel   and see how rotations works
21:12 RealBadAngel   https://github.com/RealBadAngel/facedirtools
21:13 RealBadAngel   to see right direction of rotations player shall look in the very same direction
21:13 RealBadAngel   i mean look dir = axisdir
21:14 RealBadAngel   theyre all counter-clock wise
21:15 RealBadAngel   but when you will look at face you will see it clock wise
21:16 RealBadAngel   PilzAdam, use mode 3 of screwdriver to change axisdir
21:16 RealBadAngel   then set it to mode 1 to change rotations
21:18 RealBadAngel   in console it will be dumped in format (axisdir,rotation)
21:23 celeron55      thexyz: that is true; but RBA didn't provide a specification for how dir_to_facedir should be extended
21:24 RealBadAngel   the function needs to be rewritten
21:24 RealBadAngel   shall i do that?
21:25 celeron55      well, directly modifying the parameters of a node can't be the official API method of doing the thing
21:25 RealBadAngel   of course
21:27 RealBadAngel   but im not quite sure how the placement prediction should be done
21:27 RealBadAngel   i will try to write some and test it to find best solution
21:29 RealBadAngel   but anyways the screwdriver shall become a default tool
21:29 VanessaE       agreed.
21:29 RealBadAngel   its way too handy, were already for old facedir
21:29 RealBadAngel   now it is absolute must-have
21:30 celeron55      we can't include a modern screwdriver in the game, as the main game's world is situated in such a point of time that screwdrivers simply didn't exist
21:31 celeron55      or, well
21:31 RealBadAngel   so lets call it crowbar
21:31 VanessaE       celeron55: what IS the time period you're trying to model anyway?
21:32 celeron55      it might also be necessary (or actually, is) that the tool is logically useless in survival gameplay
21:32 celeron55      VanessaE: i don't really know
21:32 VanessaE       because screwdrivers have been around since the late 15th century\
21:32 RealBadAngel   cmon, we do have already force fields and stargates in game ;)
21:33 celeron55      no we don't, they are independent mods
21:33 thexyz         VanessaE: well, it obvious, screwdrivers don't exist yet mese does
21:33 thexyz         VanessaE: can't you remember?
21:33 VanessaE       heh
21:34 RealBadAngel   celeron55, ofc in mods, but we do
21:34 celeron55      i don't know where to set the aim here
21:34 RealBadAngel   so maybe rename it to crowbar?
21:34 RealBadAngel   its not so modern lookin :)
21:35 celeron55      okay let's say it can be a screwdriver, but it cannot look like plastic and galvanised metal
21:35 VanessaE       crowbar to flip the object over, screwdriver to rotate it around the current Y axis.
21:35 VanessaE       that's how I'd do it
21:35 RealBadAngel   you dont like modes idea?
21:37 VanessaE       some users may get confused by that
21:37 RealBadAngel   celeron55, have you tried screwdriver?
21:37 PilzAdam       celeron55, my plan how to handle such requests is to just add it to the "Build" game, and let survival be the good game :-)
21:37 celeron55      i'd say default's world is basically "a human being having been thrown in the middle of nowhere in 1850
21:37 celeron55      "
21:37 VanessaE       celeron55: that sounds fair enough.
21:38 VanessaE       that's "wild west" territory in US-speak ;-)
21:38 RealBadAngel   so 350yrs after screwdriver was invented
21:38 celeron55      or maybe 1800
21:38 celeron55      or 1750 - but something like that
21:39 PilzAdam       https://github.com/PilzAdam/common/commits/stairs
21:39 PilzAdam       stairs and slabs that use 6d facedir
21:40 PilzAdam       but I dunno about this, it feels rather hacky
21:40 VanessaE       PilzAdam: don't forget to convert whatever is already in the map
21:41 VanessaE       (so that you can delete the extraneous nodes)
21:41 PilzAdam       the upside_down nodes are still registered
21:41 RealBadAngel   imho no need to
21:41 PilzAdam       and its the best solution IMO, since ABMs are slow
21:41 RealBadAngel   all extra nodes drops one default
21:42 VanessaE       RealBadAngel: final report on memory:  without item preload enabled, I seem to hover around 6.3GB or so with 512px.  with it enabled, about 7.6Gb or so seems about right.  Before 6d, I guess I used about 400MB more than now.
21:42 RealBadAngel   so user will have just to pick them up and place again
21:42 VanessaE       PilzAdam: abms only trigger if there is a node to trigger them with.
21:42 VanessaE       so after all stairs in an area have been translated, the abm has nothing left to operate on, so it doesn't run
21:43 celeron55      PilzAdam: wouldn't after_place_node be useful for that?
21:43 RealBadAngel   400mb less for a change that actually adds functionality is quite good, dont ya think? :)
21:43 VanessaE       RealBadAngel: quite so.  that's just a rough guess based on past numbers, but it seems right to me.
21:43 PilzAdam       celeron55, no, you dont get pointed_thing there
21:47 hmmmm          hah.  does it actually matter? just leave the screwdriver in there
21:47 hmmmm          i like it
21:48 RealBadAngel   how many uses shall it have?
21:49 PilzAdam       add it to "Build", but with a wooden grip
21:50 VanessaE       RealBadAngel: as many as it has now, even in the old days a screwdriver would still last a while
21:50 RealBadAngel   hehe by now its infinite :)
21:50 VanessaE       I mean as many uses as the one in technic
21:51 VanessaE       (that's the one I'm used to)
21:51 RealBadAngel   ok
21:51 PilzAdam       VanessaE, what about the bones mod in survival: https://github.com/minetest/survival/pull/1 ?
21:51 RealBadAngel   btw, it appears that Jeija have code for my portals idea for 2 months already
21:52 VanessaE       PilzAdam: if I were playing survival, I think I'd be mildly against that, but I won't stand in your way on it.
21:52 VanessaE       since I play in creative these days
21:52 celeron55      ...why is PilzAdam asking survival stuff from VanessaE?
21:52 PilzAdam       whom should I ask else?
21:53 celeron55      good question
21:53 celeron55      well, you can ask me 8)
21:53 PilzAdam       IIRC you wanted the bones mod to be added?
21:54 celeron55      i say throw it in; if it turns out not working, it can be removed
21:54 RealBadAngel   https://github.com/Jeija/minetest/tree/rejoin
21:54 RealBadAngel   have you guys seen that?
21:54 Exio           i think the items should "dissapear" after some time
21:54 celeron55      we can have a nice period of just throwing stuff in and seeing if it works now - feel free to take advantage of it
21:54 PilzAdam       RealBadAngel, yes
21:54 celeron55      (for survival and build)
21:55 Exio           so if you didn't get it in like... 2 hours? (configurable) you just lost the stuff
21:55 RealBadAngel   what do you think about it?
21:56 PilzAdam       Exio, after 20 minutest the items are opened to the public
21:57 Exio           i saw
21:57 Exio           but imo the stuff should not be "permanent" in the ground or so
21:57 PilzAdam       hm...
21:57 PilzAdam       dunno
21:58 Exio           if not, you got killed in lava
21:58 VanessaE       PilzAdam: that one bug about the bones node occasionally replacing things when it appears, has that been fixed?
21:58 Exio           and you take your time for removing it node per node
21:58 Exio           2 days later you get your stuff like before
22:01 PilzAdam       VanessaE, the bones replace everything
22:01 VanessaE       I mean replacing nodes in the world, i.e. it appears in place of a dirt node.
22:02 PilzAdam       only if you die inside a block
22:04 celeron55      that's not a real problem in survival
22:04 celeron55      well... okay, maybe it is, because it is exploitable in multiplayer
22:05 PilzAdam       I somone would fix that you can place nodes in yourself, then it wouldnt be a big problem
22:05 PilzAdam       *someone
22:05 celeron55      that won't make a practical difference
22:06 hmmmm          guys, about luajit, did we add BitOp to Lua yet?
22:06 hmmmm          I want luajit to be official soon
22:06 PilzAdam       hmmmm, Jeija's particlespawner is rebased
22:06 hmmmm          ah excellent
22:06 hmmmm          what was the other thing that required the network version to be bumped
22:06 PilzAdam       I tested it a bit and it seems to work
22:07 hmmmm          backface_culling?
22:07 PilzAdam       yep
22:07 hmmmm          'k, both of those then
22:07 hmmmm          will you commit that?
22:08 PilzAdam       what? the backface_culling thing?
22:08 hmmmm          yea
22:08 celeron55      what is this backface_culling thing?
22:08 hmmmm          isn't that what it's called?
22:08 Exio           hey, stop adding features and stuff, i can't finish compiling the last git without having to do other git pull :(
22:08 Exio           :P
22:09 hmmmm          this https://github.com/doserj/minetest/commit/75fe6d3e0c7d81405b833a476ca196e2c24eb8ab
22:09 PilzAdam       https://github.com/doserj/minetest/commit/75fe6d3e0c7d81405b833a476ca196e2c24eb8ab
22:09 PilzAdam       oh
22:10 celeron55      if you bump the protocol version, there's also the rightclickable nodedef field waiting to be placed in a final position in a serialized nodedef (altough it might just as well be at the end like it is now)
22:12 celeron55      i really should figure out the final form of the BinaryKeyValueList thing in the meta_set_nodedef branch, it'd allow much more robust and flexible serialization
22:12 hmmmm          rightclickable nodedef?
22:12 hmmmm          i've heard that phrase before but i have no knowledge about this at all
22:13 celeron55      just a boolean nodedef field
22:13 hmmmm          what, does that trigger some lua event on right click?
22:14 celeron55      it tells the client it's supposed to inform the server in a raw way if the player rightclicks a node instead of placing something or expecting a formspec
22:14 celeron55      this is the kind of lua/network stuff you probably don't want to care about :P
22:15 hmmmm          "don't want to care about" - i like how that's phrased because it's so true
22:16 hmmmm          i care about too many things already, any more and i'll lose quality for quantity
22:16 * celeron55    knows the basics of management
22:17 PilzAdam       https://github.com/PilzAdam/minetest/commits/test
22:17 PilzAdam       everything fine?
22:17 PilzAdam       (btw: nice branchname, isnt it? :-))
22:17 RealBadAngel   http://i.imgur.com/3JRscCQ.png
22:17 RealBadAngel   any better?
22:18 hmmmm          hah, that positioning makes me feel like i'm going to stab somebody with that screwdriver
22:18 PilzAdam       RealBadAngel, what resolution is it?
22:19 celeron55      RealBadAngel: do not make the rotation values behave differently regarding to the face direction - if that is what you have done
22:20 PilzAdam       is the branch good to merge?
22:20 troller        https://github.com/minetest/minetest/pull/555/files - mapgenindev larger far biomes + floatlands + tunes caves
22:20 RealBadAngel   afaik theyre all clockwise when you look at the top face
22:20 celeron55      RealBadAngel: it must behave in a constant way; think of a game on minetest where you move freely in 3D space with no top or bottom or left or right or anything
22:21 celeron55      RealBadAngel: you do not want inconsistent behavior on that; it would be suicide
22:21 celeron55      suicide design
22:21 celeron55      it's a new term for this!
22:21 RealBadAngel   it took us 3 days (nights rather) me and khonkortisian to test it
22:21 hmmmm          that design makes me want to commit sudoku
22:21 RealBadAngel   lol
22:23 RealBadAngel   celeron55, ive decided to take top face as pointer. it always points the axisdir
22:24 RealBadAngel   use screwdriver mode 4 to bring top face in pointed one
22:24 RealBadAngel   and then rotate it
22:24 VanessaE       it sounds like the node is rotated around the Y axis first, then the top is re-oriented to point to whatever face the extended facedir says
22:24 RealBadAngel   nothing is reorientated
22:25 RealBadAngel   old facedir is case when top points y+
22:25 RealBadAngel   top of a chest, or furnace or whatever
22:25 celeron55      i want it so that if you look a node at the top face side and rotate it, it will always rotate eg. counter clockwise independently of what axis the top is actually pointing in world coordinates (i think this is what you now have)
22:25 hmmmm          hm just a side question.. why is sandstone dirtlike?
22:26 RealBadAngel   all are counter-clock wise, as old facedir was
22:26 RealBadAngel   but remember when you will look at rotated face your view dir is different
22:26 RealBadAngel   so you will see it rotating clockwise
22:27 RealBadAngel   to see real rotations you would have to stand behind node
22:27 PilzAdam       hmmmm, all the 3 commits about the protocoll version: https://github.com/PilzAdam/minetest/commits/test is it good to merge?
22:27 celeron55      PilzAdam: i think it's probably good (just had time to look through these)
22:28 hmmmm          yeah it looks good
22:29 RealBadAngel   PilzAdam, screwdriver has modes displayed on it (they have to be in greater resolution to be readable and lookin nice) so the screwdriver is upscaled to 128x but with bigger pixels. So it looks like it was 16x
22:30 VanessaE       RealBadAngel: alternative:  display a colored square in one corner to indicate the mode.  red->yellow->green->purple maybe
22:30 VanessaE       then you can get by with a true 16px image.
22:30 celeron55      or rotate the screwdriver? 8)
22:30 RealBadAngel   lol
22:31 RealBadAngel   upside down would be lookin like you are trying to stab yourself with it :)
22:31 VanessaE       or a colored band on the handle is better
22:32 RealBadAngel   it shows messages anyway
22:33 PilzAdam       since the backface_culling works now, should new_style_water be true by default?
22:33 RealBadAngel   but i like that idea with modes, i can configure my tools and keep more preconfigured tools on hotbar and switch
22:34 celeron55      PilzAdam: of course now; it's looks bad and performs bad
22:34 celeron55      not*
22:34 PilzAdam       heh, nice typo :-)
22:36 Exio           so, will someone update http://dev.minetest.net/Changelog#0.4.5_.E2.86.92_.3F ?
22:37 PilzAdam       yes
22:38 RealBadAngel   so what about screwdriver? different coloured handle or keep with modes number?
22:39 Exio           i'd say the color
22:39 Exio           but i'm just an user!
22:39 RealBadAngel   numbers are easier to remember as for me
22:40 celeron55      leave the number for now
22:50 PilzAdam       Exio, done
22:57 hmmmm          hmm
22:58 hmmmm          pilzadam, what do you think about making furnaces a light source while active?
22:58 hmmmm          how hard would that be in lua?
22:58 Exio           one line
22:58 VanessaE       I thought they already were>
22:58 VanessaE       ?
22:58 PilzAdam       they are
22:59 PilzAdam       hmmmm, buy some glasses
22:59 hmmmm          they are?wtf hold on
22:59 PilzAdam       init.lua:1547
22:59 PilzAdam       (of common/default)
23:00 hmmmm          ahh
23:00 hmmmm          man why did i not see this
23:01 Exio           hahaha
23:06 PilzAdam       aw, man, I want to push something
23:07 hmmmm          reminds me of that static-X song, "push it"
23:08 RealBadAngel   http://realbadangel.pl/screwdriver.zip
23:08 RealBadAngel   can somebody try it?
23:30 rarkenin_      Given a name, is it possible to get a peer id within C++? How id this done?
23:32 RealBadAngel   hold on
23:37 RealBadAngel   getPlayer(name)->peer_id)
23:39 celeron55      Segmentation Fault
23:40 RealBadAngel   celeron55, due to what?
23:42 hmmmm          because getPlayer can return null if name isn't found
23:42 hmmmm          perhaps
23:43 RealBadAngel   https://github.com/minetest/minetest/blob/master/src/scriptapi.cpp#L790
23:43 RealBadAngel   there is example
23:44 rarkenin_      OK.
23:46 RealBadAngel   so what about screwdriver? PilzAdam?
23:46 PilzAdam       its fine for "Build" game, I guess
23:47 RealBadAngel   imho it shall go for all the modes...
23:47 RealBadAngel   hmmmm, celeron55?
23:47 PilzAdam       no, it definetly doesnt go into survival
23:48 hmmmm          i didn't try the screwdriver
23:48 hmmmm          don't really have an opinion of it, i think it's fine though
23:48 RealBadAngel   so try :)
23:49 hmmmm          busy
23:49 hmmmm          i'll try it when i grab the latest minetest_game and other things
23:49 PilzAdam       what does a screwdriver add to survival gameplay?
23:49 RealBadAngel   ability to rotate facedired nodes
23:49 celeron55      PilzAdam: i agree that for real survival, it's useless
23:50 celeron55      but it's useful for many build-focused games so put it in common
23:50 RealBadAngel   not really
23:50 RealBadAngel   it allows to rotate things like wires, delayers, all discrete stuff
23:51 RealBadAngel   building circuits with it is much easier
23:52 celeron55      the point of having these different games is that we can make them *actually* focus on different things; the build game is going to be much closer to minetest_game than the survival one will be
23:52 RealBadAngel   also truth
23:53 RealBadAngel   so let it be in build then
23:53 RealBadAngel   PilzAdam, will you merge it?
23:54 PilzAdam       in wich mod should it be?
23:54 celeron55      or, actually, i don't even know if the last sentence is correct
23:54 RealBadAngel   its standalone mod
23:54 celeron55      i don't have enough experience of the building part of minetest that i could know
23:54 PilzAdam       celeron55, I agree to that sentence
23:55 PilzAdam       the building part is about a lot of funny blocks to place
23:55 celeron55      PilzAdam: leave it as a separate mod; there is no reason not to
23:55 PilzAdam       RealBadAngel, then add it to common
23:55 RealBadAngel   celeron55, believe me this tool is extremaly useful. it was even before 6d facedir
23:57 RealBadAngel   building out of small parts like microblocks, slabs, panels with screwdriver is just easy