Time  Nick           Message
01:23 hmmmm          paramat, looks good but I noticed I spawned inside of the ground
01:24 hmmmm          that's an indication that your getGroundLevelAtPoint() function is off
01:24 hmmmm          mind you, it does not need to be perfect, but it should be at least close to correct
01:40 paramat        sorry i havent written getGroundLevelAtPoint() yet, now of course i realise it is needed for spawn player
01:46 Zeno`          Is it possible to reopen a closed PR? If so, will it then pick up the latest changes from my repo (it doesn't seem to have while it's closed). Or, is it better to open a new PR. See https://github.com/minetest/minetest/pull/1824
02:01 Zeno`          hmmmm, are you there?
02:02 Zeno`          Maybe it's best to make a new PR?
02:24 paramat        thanks for the merge O/
02:27 kaeza          https://mediacru.sh/P9jTwnwRBq6m
02:27 kaeza          oops, wrong chan
02:27 shadowzone     Sweet
03:13 Zeno`          kahrl, the issue RBA was talking about last night exhibits itself when using directX9 as well, so maybe it's to do with the version of opengl being used
03:15 Zeno`          http://i.imgur.com/EtIX9AS.png
03:25 kahrl          Zeno`: oh... yeah that could be
03:43 RealBadAngel   kahrl, http://pastebin.com/ppWGYubX
04:32 hmmmm          Zeno`, yeah it's better to create a new PR
05:56 hmmmm          Zeno`
05:57 hmmmm          yeah it's a better idea to make a new PR
05:57 Zeno`          I did that in the end. I shouldn't have closed it in the first place, but *shrug*
05:59 Zeno`          You may as well look at it now :P ShadowNinja looked the other day and nothing has changed, but the more feedback the better. #1825
05:59 ShadowBot      https://github.com/minetest/minetest/issues/1825 -- Create faster key cache for main game loop (client) by Zeno-
05:59 hmmmm          christ that's a huge difference
05:59 Zeno`          10x
06:00 Zeno`          and since it's in the main loop I think it's worth it
06:00 hmmmm          something like that is always worth it
06:00 hmmmm          the "oh it's fast enough" attitude in minetest is what got us in this state to begin with :(
06:01 hmmmm          you might take a code path that's 50% slower, but that's okay because it's not speed critical, but you do that everywhere and you end up with a program that's 50% slower everywhere and you can't do anything about it
06:01 hmmmm          inefficiencies being baked into the program like that, not too good
06:03 Zeno`          it's only a small change as well :)
06:03 Zeno`          small as in method, not result
06:03 hmmmm          indeed..
06:04 hmmmm          settings could be a lot faster if a good synchronization scheme were used
06:04 hmmmm          contention is extremely rare in the write case, whereas there are many readers
06:04 hmmmm          it can be optimized
06:05 Zeno`          I've thought of that as well. Maybe post 0.4.11 it could/should be looked at. I mean maps/hashes are fast and O(1) but they're not the same O(1) as an array lookup (i.e. they have to iterate over the strings to compute the hash)
06:05 hmmmm          nope
06:05 hmmmm          std::map is O(log(n)
06:05 Zeno`          oh is it?
06:05 hmmmm          yup
06:05 Zeno`          well there you go hehe
06:06 hmmmm          you might be thinking of std::unordered_map which is C++11
06:06 Zeno`          I thought it was O(1) :/
06:06 Zeno`          I must be
06:06 hmmmm          std::map is an RB tree
06:06 hmmmm          elements are ordered lexographically in the case of std::string
06:06 Zeno`          yeah I see that now (looking at the docs)
06:07 Zeno`          ok  std::unordered_map  O(1) is still not (quite) as fast as array lookup; it cannot be
06:07 hmmmm          of course not
06:07 hmmmm          O(1) is the optimal case
06:07 Zeno`          c++ std is too large!
06:07 hmmmm          worst case is 'guaranteed' to be 'at most O(n)'
06:07 Zeno`          I need the abridged version heh
06:07 hmmmm          by the standard
06:08 hmmmm          you can do a guaranteed O(2) hashtable
06:08 Zeno`          well, that's not too bad
06:09 hmmmm          cuckoo collision resolution
06:09 Zeno`          how does it guarentee that? I have to look .. oh ok
06:09 Zeno`          never heard of it
06:09 * Zeno`        googles
06:10 Zeno`          Oh I've seen that. Didn't know that's what it was called
06:10 hmmmm          i think it's pretty clever
06:10 hmmmm          i love hashtables a lot but they're really hell on the cache
06:12 Zeno`          yeah. in many instances they're perfect though
06:13 Zeno`          not here though :)
06:26 RealBadAngel   hmmmm, <celeron55> is there opposition in having Zeno` as a core developer?
06:26 hmmmm          no
06:27 RealBadAngel   so its like votes now for him being core dev
06:27 RealBadAngel   7
06:28 hmmmm          let's put it this way
06:28 hmmmm          has anybody disagreed?
06:28 RealBadAngel   no
07:04 RealBadAngel   kahrl, http://i.imgur.com/ByfSbnm.png
07:05 RealBadAngel   i have almost fixed the issues with wielded (and added there shaders)
07:20 hmmmm          wooo...
07:20 hmmmm          i have the managers implemented
07:20 hmmmm          now to just change the mapgens over to use those instead of directly accessing emergemanager and then i have to do all the lua api junk
07:20 hmmmm          shoot me now
07:35 Zeno`          what's that image RealBadAngel?
08:00 Zeno`          Can #1810 be merged?
08:00 ShadowBot      https://github.com/minetest/minetest/issues/1810 -- Fix profiler values not being updated (F6) and not being logged by Zeno-
08:01 hmmmm          sure
08:01 hmmmm          just FYI, it's one tab for a function declaration continuation line
08:01 hmmmm          not two
08:02 hmmmm          also, doesn't getting the profiler print interval every frame go against your philosophy of caching settings?
08:04 Zeno`          yes it does. But it's not a "key"
08:04 Zeno`          there are similar issues in camera.cpp
08:04 Zeno`          I figured I'd tackle those post-0.4.11
08:05 Zeno`          mainly because it will probably mean larger changes to g_settings
08:05 Zeno`          Thanks for the line continuation clear up. I think it was a source of confusion not just for me
08:06 hmmmm          some random person with a random ip address added that rule to the code style guidelines
08:06 hmmmm          lol
08:06 Zeno`          profiler print interval can be changed using /set I think
08:06 Zeno`          that's one of the reasons I was wary about it
08:06 hmmmm          ahh
08:06 hmmmm          probably a good idea then
08:07 Zeno`          a random rule! great :)
08:08 hmmmm          all the big projects seem to agree on one tab for a function declaration continuation line for parameters
08:09 Zeno`          well, that's how I *used* to do it (and do it in my own projects)
08:10 Zeno`          Noted :)
08:24 Zeno`          I can't merge yet. I guess the vote (or whatever it is) is still going on =)
08:24 Zeno`          bbiab (dinner)
08:39 hmmmm          or perhaps celeron simply isn't around at the moment because time zones
08:40 hmmmm          he's the only one with the permissions necessary to add/remove developers
08:42 Zeno`          oh yeah
08:48 celeron55      Zeno`: i added you now
08:48 celeron55      (just woke up)
08:52 Zeno`          thank you. Back to dinner :)
09:16 Zeno`          umm, why did I just do 2 commits?
09:16 Zeno`          dammit... my first merge and I stuffed up
09:17 Zeno`          Should I not have deleted the branch?
09:18 Amaz           Should this pull request be closed, as the commit it contains seems to have been pushed? https://github.com/minetest/minetest/pull/1818 https://github.com/minetest/minetest/commit/10a47b7feef3278792c567f5b954ea4c3712319c
09:20 Zeno`          I think it should be, but I won't close it until I know what I did wrong with my last merge heh
09:23 Zeno`          I have no idea what has happened
09:23 Zeno`          help!
09:23 Zeno`          RealBadAngel ^^^
09:36 Zeno`          shit, I'm sorry. I need to know how to stop this happening again
09:49 kaeza          Zeno`, use `git rebase`
09:50 Zeno`          there was only one commit
09:50 Zeno`          I don't get it :/
09:51 Amaz           Do you need to merge it locally with --no-commit, and then push to master?
09:54 Zeno`          I dunno. I just pressed the button on github (obviously wrong lol)
09:58 Amaz           From http://dev.minetest.net/Git_Guidelines Use rebase, not merge, to get linear history. [3] [3] For a github pull request, this is easiest done by appending .patch to the pull request URL, wgetting it to your project directory and doing git am whatever.patch. Similarly for single commits.
10:00 Zeno`          ok cool
10:01 Zeno`          I made a mistake and it looks like that's how to stop it again... thanks
10:01 Zeno`          that actually makes sense
10:01 Zeno`          so I sign it off rather than merge
10:03 Zeno`          yep, cool
14:41 Zeno`          kahrl, the line notes make sense. I was trying, as much as possible, to keep the commit with the same logic as the original main.cpp. Should I update them?
14:42 Zeno`          I probably should
14:43 kahrl          ah
14:44 kahrl          maybe do it in a later commit then
14:44 kahrl          (also I just started commenting, so more to come :P)
14:44 Zeno`          thanks :)
14:45 Zeno`          I'm actually asleep right now so will look more closely tomorrow
14:46 kahrl          sleeptyping?
14:47 Zeno`          yeah... happens
14:47 kahrl          just don't do this: http://www.americanbanker.com/people/bank-worker-in-germany-falls-asleep-transfers-millions-1059834-1.html
14:48 Zeno`          well... wouldn't be so bad if it was to my account
14:49 Zeno`          lol though
16:23 RealBadAngel   hi guys
16:23 RealBadAngel   kahrl, on?
16:26 kahrl          RealBadAngel: slightly
16:27 RealBadAngel   i was working on the issue
16:27 RealBadAngel   i got it almost working with shaders support added
16:27 kahrl          any ideas? I see other people have been getting it now too
16:27 RealBadAngel   definition of the material in extrude wasnt enough
16:28 kahrl          I see
16:28 RealBadAngel   i run set material for each buffer in setitem
16:28 RealBadAngel   and found this place nice to apply shaders too
16:28 kahrl          so maybe it depends on the irrlicht version?
16:28 RealBadAngel   i just need to turn ofg (convert) waving materials
16:29 RealBadAngel   *off
16:29 RealBadAngel   idk
16:29 RealBadAngel   but materials got overwritten at some point
16:29 kahrl          hmm
16:29 RealBadAngel   so i decide to set the materials no matter what
16:29 RealBadAngel   and it works
16:29 kahrl          sounds good
16:30 RealBadAngel   have you saw screenshots i posted here?
16:30 kahrl          I saw some, not sure if I saw all
16:30 RealBadAngel   http://i.imgur.com/ByfSbnm.png
16:30 RealBadAngel   32x textures, auto normalmaps
16:31 kahrl          nice
16:31 RealBadAngel   http://i.imgur.com/VK2588Y.png
16:31 RealBadAngel   256x normalmaps with parallax mapping
16:32 kahrl          what's up with the black dots and the lines on the edges? are they meant to be there?
16:32 RealBadAngel   no
16:32 RealBadAngel   i saw that artifact too
16:32 RealBadAngel   filtering reduces that but not eliminates
16:33 RealBadAngel   once i had all the cases working i will try to find out the reason
16:33 RealBadAngel   propably rounding errors
16:34 RealBadAngel   but that we can call fine tuning just
16:35 RealBadAngel   please do note that we had similar problems before, with dirt with grass
16:35 kahrl          why that node in particular?
16:35 RealBadAngel   on some step angles grass from top was seen at the bottom
16:36 RealBadAngel   so you were able to see thin green lines
16:36 kahrl          with anisotropic filtering on?
16:36 RealBadAngel   with and without
16:36 kahrl          strange
16:37 RealBadAngel   that may be engine bug, not ours
16:37 RealBadAngel   but still
16:37 kahrl          I had those artifacts on wield meshes in early versions of wieldmeshnode
16:37 kahrl          thought they were gone
16:37 RealBadAngel   nah, theyre still here
16:38 RealBadAngel   hold on a sec
16:38 kahrl          I really hope they can be fixed
16:38 Zeno`          floating point calcs are not commutative so order is pretty important
16:39 RealBadAngel   http://pastebin.com/c0UsREqn
16:39 RealBadAngel   WIP of setItem
16:39 Zeno`          e.g. x * y is not necessarily the same as y * x
16:39 RealBadAngel   you will be able to see the different approach
16:40 RealBadAngel   i serve the cases first, then work with materials, no matter the case used
16:41 kahrl          RealBadAngel: I think it would be easier to move the material handling code to the end up changeToMesh
16:41 kahrl          then you won't need to change any other function and it automatically works with setCube and setExtruded as well
16:42 RealBadAngel   propably yes, i will see
16:42 kahrl          oh, but you are accessing ContentFeatures
16:43 RealBadAngel   ah yes, i need drawtype
16:43 RealBadAngel   and tiles and their shaders
16:43 kahrl          do everything except the part that needs f.tiles in changeToMesh
16:43 kahrl          and everything that needs f.tiles in setCube
16:44 kahrl          since you have access to TileSpec[6] in there
16:45 RealBadAngel   ok
16:45 hmmmm          hah
16:45 RealBadAngel   btw, https://github.com/minetest/minetest/pull/1808
16:45 hmmmm          jeez don't you guys have jobs
16:45 RealBadAngel   i do, a new one :P
16:46 RealBadAngel   now im IT specialist
16:46 kahrl          I have a part time job ;)
16:46 RealBadAngel   with lotsa time to code stuff
16:48 kahrl          the rest of the time I should be studying but I don't want to today
16:48 RealBadAngel   about 1808, it was ok but guy merged new commits into his branch, so it needs to be extracted
16:48 RealBadAngel   but its a bugix and i want it to be merged
16:48 RealBadAngel   can i merge the original commit?
16:49 kahrl          does it break any mods?
16:49 RealBadAngel   no
16:49 RealBadAngel   checked that
16:49 RealBadAngel   i was worried bout mesecons, but not
16:49 kahrl          but others that do use rotate_node?
16:50 RealBadAngel   so far i havent found any that could be affected
16:50 kahrl          homedecor maybe?
16:50 RealBadAngel   if so, safe
16:50 VanessaE       homedecor's lights use that
16:50 RealBadAngel   idk who made those rotations
16:51 VanessaE       I did. :P
16:51 VanessaE       shadow rewrote it though
16:51 * RealBadAngel cast a curse on both VE and SN :P
16:51 VanessaE       I didn't, however, have anything to do with 1808.
16:51 RealBadAngel   those tables were obviously wrong
16:52 RealBadAngel   just this guy stepped over it
16:52 RealBadAngel   this is very similar to wallmounted being upside down for 0 value
16:53 RealBadAngel   thus breaking any conventions
16:53 VanessaE       idk what he's trying to do but the rotation was done such that the textures would align with non-facedir nodes
16:53 VanessaE       the original rotations that is
16:53 RealBadAngel   facedir handles the rotations
16:53 RealBadAngel   some have DECIDED to use such 6d rotations not others
16:54 VanessaE       yeah he just doesn't understand the purpose of that routine
16:54 RealBadAngel   he just apply extra 90 degs rotation
16:54 RealBadAngel   propably after modifiers used in textures
16:55 RealBadAngel   and forgot that engine doesnt need it anymore
16:56 RealBadAngel   then never used this feature and it was waiting just for some1 who needed it
16:56 VanessaE       RealBadAngel: check it with stone brick stairs
16:56 kahrl          ShadowNinja: comments? ^^
16:56 VanessaE       see if those are still properly aligned
16:56 RealBadAngel   doesnt apply
16:56 VanessaE       those were my test case
16:56 RealBadAngel   we are talking bout wallmounted only
16:57 RealBadAngel   stairs are not wallmounted
16:57 VanessaE       um, no
16:57 VanessaE       his change is to minetest.rotate_node()
16:57 RealBadAngel   ah i see
16:57 VanessaE       that'll affect stairsplus
16:57 VanessaE       moreblocks
16:57 VanessaE       which affects fuckloads of mods
16:57 RealBadAngel   ok i will see
16:58 RealBadAngel   but only placing of the node
16:58 RealBadAngel   eg giving placed node correct param2
16:58 VanessaE       but if stone brick slabs et al. still aligns properly next to a stone brick block then it's probably fine
16:58 VanessaE       yeah
16:58 VanessaE       that's all he's doing here
16:59 RealBadAngel   i will check the stairs and then we will see
16:59 RealBadAngel   ok
17:10 Calinou        when does minetest_game get some biomes by default, for use with v5 and v7?
17:10 Calinou        this would make the mapgens usable by default
17:11 VanessaE       Calinou: paramat is going to make a pull request.
17:11 Calinou        what biomes will it have?
17:11 Calinou        I'd rather have less than 25 biomes, with less than 25 custom nodes, less than 250-block high mountains
17:11 Calinou        if you know what I mean…
17:11 * Calinou      rolls eyes
17:14 Calinou        biomes of paragenv7 just don't look good and are too complex
17:14 VanessaE       I would assume the half a dozen or so standard ones
17:14 VanessaE       sand, grass, etc.
17:15 VanessaE       have "we" settled on what exactly should happen with snow?  as in where it should be by default?
17:15 VanessaE       (far north/south, or just all over/patchy like with snow mod)
17:16 Calinou        forest + plains + desert + snow
17:16 Calinou        is pretty much all we need
17:16 Calinou        would be like v6, but with snow biome (to use the new nodes)
17:16 VanessaE       beaches, lakes/oceans?
17:16 Calinou        that too
17:16 Calinou        but certainly not 6 types of forest, 4 types of plains, with 3 different grass colours
17:16 Calinou        doesn't fit minetest_game at all
17:16 VanessaE       no, you're right about that
17:17 VanessaE       but I think one more is needed
17:17 VanessaE       maybe that "dried out" one from mg?
17:17 Calinou        and of course, the grass texture should not be applied to all sides (like it is in paragenv7 currently)
17:17 VanessaE       I forget what it's called
17:17 VanessaE       (obviously I don't mean desert)
17:17 Calinou        a taiga (+ pine trees) would be OK
17:17 Calinou        but it needs to stay relatively simple
17:17 VanessaE       that would work too
17:17 VanessaE       just ONE more, not a dozen
17:19 kilbith        "forest + plains + desert + snow [...] is pretty much all we need"
17:19 kilbith        ^ agreed
17:48 RealBadAngel   my opinion on biomes: we need a reason for a player to explore the world
17:48 RealBadAngel   what for nearly unlimited one when its no difference when you stick to spawn or go 20k in any direction?
17:49 RealBadAngel   bring on the damn variety to the world
17:50 RealBadAngel   atm fly around in 200-300 nodes radius and you have saw all
17:50 VanessaE       then said variety needs to be dependent on location itself, yes
17:51 RealBadAngel   maybe land shape will be different, but thats all
17:51 RealBadAngel   absolutely no point for regular player to explore the world
17:51 RealBadAngel   that kills gameplay
19:05 celeron55      having things look different but be functionally the same doesn't help though
19:06 celeron55      needs actual new content
19:06 Calinou        too much biomes requires players to explore too much
19:06 Calinou        == bad for gameplay, requires players to be jobless
19:06 Calinou        digging stuff is too slow already
19:08 celeron55      now, where's our game designer to fix these issues
19:10 celeron55      well hopefully there is one working on minetest_game 8-)
19:11 * celeron55    irresponsibly delegates
19:13 Calinou        BlockMen obviously
19:13 Calinou        (sarcasm)
19:14 Calinou        celeron55, who can send DMCAs about the illegal apps?
19:14 Calinou        just you, or can other core devs do it?
19:20 VanessaE       in theory any dev whose code is still in Minetest could.
19:21 VanessaE       (but you need to be able to prove there's a problem, prove it's your code, blah blah blah blah)
19:21 Calinou        I have 5 lines!
19:21 * Calinou      hops on a boat
19:25 RealBadAngel   celeron55, with exploring i mean lookin for rare resources, new kinds of building materials etc, specific to those biomes
19:25 RealBadAngel   be it on the ground level or deep below
19:26 RealBadAngel   those two things have to be bound together, new biomes with new materials
19:26 Calinou        adding a snow biome + taiga with trees is a good addition already
19:26 Calinou        better than what we have currently: stone everywhere.
19:28 RealBadAngel   give taiga trees ability to craft better chests for example
19:28 RealBadAngel   that would be a reason to look for those biomes
19:28 Calinou        makes no sense
19:28 Calinou        absolutely none
19:29 Calinou        why would taiga wood be “better” in any way
19:29 Calinou        it's at best inconsistent, at worst senseless
19:29 RealBadAngel   better wood, better chests
19:29 RealBadAngel   just an example (or idea) something to make this biome worthy
19:29 Calinou        come with consistent ideas
19:29 Calinou        eg. special flower acting as good fuel
19:29 Calinou        glowing flower
19:30 RealBadAngel   ofc, that was a quick idea
19:30 Calinou        it doesn't have to realistic, but everything should make sense
19:30 RealBadAngel   but giving a picture
19:30 RealBadAngel   different look of the biome is not enough
19:31 RealBadAngel   if any biome you will be able to find group:wood having same usage, differnt kinds of wood will end up the same
19:31 Calinou        all kinds of wood should act the same, except maybe for smelting duration
19:31 RealBadAngel   now you start to make biomes the same :P
19:32 RealBadAngel   all wood, all stone all the same
19:32 RealBadAngel   and that exactly is wrong
19:33 RealBadAngel   it unifies the biomes regardless how much effort you put into them to look different
19:35 Calinou        biomes shouldn't act too different
19:35 Calinou        else you come into the problem of balance
19:35 Calinou        everyone wants to live in biome X, but no one wants to live in biome Y
19:35 Calinou        == overcrowded areas in multiplayer
19:36 RealBadAngel   aviable area is big
19:36 RealBadAngel   you can find next biome of that type
19:37 VanessaE       Calinou: then the trick is to make sure that you need things from *both* biomes
19:37 RealBadAngel   or all of them
19:37 VanessaE       RealBadAngel: indeed
19:37 VanessaE       what use is snow right now?  it has none.
19:37 Calinou        but make differerces that makes sense
19:37 Calinou        don't make wood act different like that :/
19:37 Calinou        I'd hate that
19:37 RealBadAngel   giving each biome an attribute, thing that will make look for it
19:38 RealBadAngel   why the duck you have stick to wood, it was an example
19:38 RealBadAngel   think about something other
19:39 RealBadAngel   i have no glue what it could be for each type atm
19:39 RealBadAngel   maybe snow for winter ones?
19:39 RealBadAngel   in mc snow and ice was very important, it was giving you ability to transport water into nether
19:40 Calinou        meh
19:40 RealBadAngel   and that was damn funny
19:41 RealBadAngel   and about such ideas im talking about
19:43 RealBadAngel   make player to visit mountains to get precious metals, force him to visit cloud biomes for materials to build flying stuff
19:43 RealBadAngel   make him go deep to find lava to fuel them
19:44 RealBadAngel   MAKE HIME EXPLORE THE WORLD :)
19:45 RealBadAngel   otherwise you can stick that +- 31k deep where no one can see ;)
19:46 RealBadAngel   nobody will check if thats true
20:03 kilbith        [Bug] latest -dev (30min ago) : "rmdir errno 39 : folder is not empty" shows in chat & debug when I delete a world
20:56 kahrl          serverlist.cpp sends g_settings->get("mg_name") when it probably should send mg_name from map_meta.txt
20:56 Amaz           Ah.
21:08 kahrl          Amaz: could you test #1828?
21:08 ShadowBot      https://github.com/minetest/minetest/issues/1828 -- Serverlist: announce mg_name from map_meta.txt instead of minetest.conf by kahrl
21:09 Amaz           Okay.
21:19 Amaz           Server running that branch is up.
21:19 Amaz           And it works!
21:22 kahrl          nice, thanks
21:23 Amaz           No problem!
21:23 Amaz           Thanks for the quick response to the bug report!
21:24 kahrl          any devs around to agree with #1828?
21:24 ShadowBot      https://github.com/minetest/minetest/issues/1828 -- Serverlist: announce mg_name from map_meta.txt instead of minetest.conf by kahrl
21:25 sfan5          kahrl: that patch qualifies as a small patch and needs no merge approval from anyone
21:25 kahrl          alright
21:25 sfan5          http://dev.minetest.net/Git_Guidelines#Rule_1_in_practice
21:26 kahrl          I know, but I usually prefer asking at least once ;)
21:31 hmmmm          booo
21:31 hmmmm          serverlist.cpp is fubar
21:32 hmmmm          the correct way to get the mapgen is to read the current working mapgenparams from the emergemanager
21:32 hmmmm          the mapgen may have been changed between when map_meta.txt was loaded and the mods run
21:32 hmmmm          also what if map_meta.txt doesn't even exist
21:32 Amaz           Well, as it is now is better than reading directly from the conf...
21:32 Amaz           As that can give wrong results.
21:33 kahrl          hmmmm: well my commit does actually do what you're suggesting if I understand correctly
21:33 hmmmm          perhaps
21:33 kahrl          I just had to keep the commit message reasonably short
21:33 hmmmm          yes that is right
21:34 hmmmm          basically, m_emerge->params.mg_name is the ONLY right way to find the mapgen
21:35 hmmmm          that's not really due to poor design, it's just that a number of separate components have control over mapgen settings before the server actually starts
22:06 paramat        so the correct way in lua to detect which mapgen is being used is using 'minetest.register_on_mapgen_init ... mgname = MapgenParams.mgname'?
22:12 paramat        hmmmm i made a start on mgv5 'getgroundlevelatpoint' but without a way to get eased 3D noise at a point it will not be accurate enough to work. i assume eased 3D noise is only available as a perlinmap currently.
22:12 kaeza          re #1820, not sure if related, but meshes as wielditems are always using mipmapping
22:12 ShadowBot      https://github.com/minetest/minetest/issues/1820 -- odd wield item glitch.
22:12 kaeza          (Intel GMA here)
22:19 RealBadAngel   kaeza,  fix for that will be here today i think
22:23 kaeza          anybody has a link to iqualfragile's mumble patch?
22:24 RealBadAngel   he promised one but i havent saw it
22:25 kahrl          https://gist.github.com/iqualfragile/bc0864e1cc9830d40e0a this?
22:26 kaeza          kahrl, thanks, that's the one
22:27 kaeza          (I think)
22:27 kaeza          I'll see if it works on windows
22:27 kahrl          you'll probably have to add an #ifdef around the #includes
22:29 RealBadAngel   could be fun
22:29 RealBadAngel   i like to have it tested
22:40 kaeza          it at least compiles (with minor modifications)
22:55 kaeza          okay, I have it running. anybody cares to test? :)
23:00 * paramat      somehow gets mgv5 spawnpoint working
23:25 Taoki          Wow. I don't wish to seem overly neative... but Minetest has such big big problems when it comes to networking x.x
23:25 Taoki          If there's a server that's even somewhat laggy, you can barely do anything. It can take a monute for even a door to open
23:25 Taoki          Or more minutes until you see anyone moving. Any command lags behind some 30 seconds.
23:26 Taoki          Maybe we should switch from UDP to TCP entirely... I don't know
23:27 proller        sctp!
23:35 Taoki          I'm going to try Freeminer again. It's safe to conclude that mainstream Minetest is not playable online... like nearly at all :/
23:38 RealBadAngel   kahrl, extrude problem is not just the black stripes
23:38 hmmmm          ahh
23:38 hmmmm          good point paramat
23:38 RealBadAngel   it is faulty somehow
23:39 Taoki          Hi RealBadAngel :)
23:39 hmmmm          trivial vix
23:39 hmmmm          fix*
23:39 paramat        i just used non eased to approximate eased =)
23:39 RealBadAngel   hi Taoki
23:40 paramat        add 8 nodes to be sure and drop player into world
23:41 RealBadAngel   http://i.imgur.com/gJp4y9a.png
23:41 RealBadAngel   http://i.imgur.com/uVqzhvT.png
23:42 VanessaE       how come I didn't see these glitches when I tested it?
23:42 kahrl          RealBadAngel: does the vertex shader change the vertex positions?
23:42 RealBadAngel   http://i.imgur.com/ZwjeZi4.png
23:42 RealBadAngel   shaders can only change vertices for nodes that have waving = 1
23:43 VanessaE       one thing I *did* see by the way is that the new code doesn't "work" with the trick HDX uses to disable extrusion on the wieldhand (black to the right of the hand, at 25% alpha)
23:43 Taoki          Is there any chance that Minetest might switch to the enet protocol too? Which is what Freeminer did, and is said to be a lot faster
23:43 RealBadAngel   and this is another issue, i do have now plants and leaves that are dancing in hand ;)
23:44 RealBadAngel   but completely unrelated
23:44 kahrl          RealBadAngel: does it happen without shaders?
23:44 RealBadAngel   lemme check for sure
23:46 kahrl          fences look fine in the hand and as dropped items for me (without shaders, of course)
23:47 RealBadAngel   oh shit
23:47 VanessaE       git HEAD (just now), without shaders or with (bumpmapping + auto normals)...  I cannot reproduce the bug
23:47 RealBadAngel   its parallax mapping responsible for that
23:47 RealBadAngel   how the fuck???
23:48 VanessaE       that explains why i can't reproduce.
23:48 VanessaE       I didn't try that :)
23:48 RealBadAngel   nobody just me have shaders on wielded ;)
23:48 VanessaE       I'm just looking at a dropped fence
23:48 VanessaE       and I STILL can't reproduce
23:48 RealBadAngel   VanessaE, i do have totally differnt code now
23:48 * VanessaE     turns EVERYTHING on
23:49 RealBadAngel   dont even try
23:49 VanessaE       ok
23:49 RealBadAngel   unless you have hacked my box and have current code ;)
23:49 VanessaE       well at least I'm consistent with kahrl then
23:50 VanessaE       thought maybe I missed something blindingly obvious when I was testing  :P
23:50 RealBadAngel   http://i.imgur.com/VK2588Y.png
23:50 RealBadAngel   i have reworked part of kahrl's code
23:51 RealBadAngel   to fix issue that some of us have and applied there shaders too
23:52 kahrl          I think the way the extrusion mesh is built is incompatible with the shaders
23:53 VanessaE       RealBadAngel: I wonder if any of this ties back in with those black lines you were getting with my slope models?  the UVs were perfect, so only way I could solve THOSE was to wrap the textures a bit.
23:53 RealBadAngel   oh cmon, it is *almost* working
23:54 RealBadAngel   somehow parallax mapping is being used on things that do not have height maps
23:55 RealBadAngel   and thats all
23:55 VanessaE       RealBadAngel: yeah, like HDX ;)
23:59 RealBadAngel   lol, now i can see textures changing on meshnodes