Time  Nick       Message
00:16 paramat    hmmmm sometime please can you review #4376 and consider something to help with the map meta hash problem? see the comments
00:16 ShadowBot  https://github.com/minetest/minetest/issues/4376 -- Mapgen: Remove unused 'flat' and 'trees' flags from mg_flags by paramat
00:44 sofar      paramat: +1 on rubenwardy
00:44 * sofar    gets back from a small vacation
00:45 paramat    cool thanks, good to have you back
01:00 paramat    i'll ask nore too, then that will hopefully be 4 approvals out of 6 and i feel enough
02:53 paramat    hmmmm, here's one that might interest you since it's decorations #4401 #4402
02:53 ShadowBot  https://github.com/minetest/minetest/issues/4401 -- Tall schematic decorations unnecessarily failing near chunk top
02:53 ShadowBot  https://github.com/minetest/minetest/issues/4402 -- Decorations: Fix decoration height check errors by paramat
06:06 hmmmm      paramat,
06:06 paramat    hi
06:07 hmmmm      i'm looking at 4376 right now and it seems good but maybe the flag constants themselves shouldn't be removed
06:07 hmmmm      instead, add a comment to the end of the lines // deprecated
06:07 hmmmm      or similar
06:07 paramat    ok agreed
06:07 hmmmm      i think it's odd that MG_FLAT was only used in mapgen_v6.cpp...
06:08 hmmmm      i think i recall each mapgen interpreting that flag on its own
06:08 paramat    afaik mgv7 did not
06:09 hmmmm      we really need to do something with the whole "flag" system
06:09 hmmmm      it's a leaky abstraction that the end user should've ever seen
06:09 hmmmm      i'm finishing up my project at work pretty soon so i should have some spare time to work on these things
06:09 paramat    nice
06:09 hmmmm      in retrospect, maybe making a hash was the compact solution but not the best solution
06:10 hmmmm      right now, a basic map_meta.txt file is like 1.5 KB of text
06:10 hmmmm      i was thinking maybe the hash setup should be changed to keeping a copy of the map_meta.txt to compare against in a new test_data/ subdirectory that goes.. somewhere
06:12 hmmmm      src/unittests/?  that would make sense in terms of keeping similar topics together, but that's for source files, so not so sure
06:12 hmmmm      i didn't expect comparing against a hash to set a barrier to so many contributors wishing to change the contents of a map meta file
06:13 hmmmm      for now just ignore the failing unit test and i'll fix it the way i described
06:14 paramat    ok. a copy to compare against sounds good
06:14 hmmmm      hell, you can even implement that if you'd like
06:14 hmmmm      i'd be around to help
06:14 paramat    krock had a problem with this hash too
06:14 hmmmm      yeah but maybe that was a good thing
06:15 hmmmm      that patch was shit
06:15 paramat    lol
06:15 hmmmm      but it's not krock's fault, it's the guy before him who changed map_generation_limit
06:17 hmmmm      yeah so i'm gonna go back to working, the sooner i can get that done the sooner i can do minetest things
06:17 paramat    O/
06:22 paramat    nore Tesseract any opinion on rubenwardy being added to mtgame team? so far +1 from me, sfan and sofar
07:04 paramat    hmmmm perhaps i should remove this UASSERT() line from the unit test so that builds do not fail on everyone's PRs? https://github.com/minetest/minetest/blob/master/src/unittest/test_map_settings_manager.cpp#L203
07:05 hmmmm      no, just surround the entire piece of code with #if 0 #endif
07:05 hmmmm      and add a comment // TODO(hmmmm or paramat): change this to compare the result against a static file
07:06 paramat    ok
07:11 paramat    nore
07:12 nore       +1 for rubenwardy as wekk (for what you asked yesterday)
07:12 paramat    ok good
07:18 paramat    celeron55 please could you consider rubenwardy as a full core dev in order to be a mtgame dev, and perhaps invite him? the most active 4 mtgame devs (out of 6 total) approve
07:18 celeron55  that's fine to me; is it fine to rubenwardy?
07:22 paramat    fairly sure, he said he's happy to help out if we need another mtgame dev
08:12 tenplus1   hi folks, would a nice dev check out pull game#1232 which fixes tnt bug
08:12 ShadowBot  https://github.com/minetest/minetest_game/issues/1232 -- fix tnt bug with huge stacks by tenplus1
10:12 juhdanad   In c++ comments should I write 'nodebox' or 'node box'?
10:14 Krock      nodebox
10:15 juhdanad   Just because the struct name is 'NodeBox'.
11:21 juhdanad   Another question: my code (in getPointedThing()) returns the normal vector of the colliding surface, which is currently v3s16.
11:22 juhdanad   But if the turning mobs could have turning selection boxes, the normal vector should be v3f.
11:22 juhdanad   Which type is preferred?
11:24 Calinou    turning selection boxes don't make much sense if the collision boxes don't rotate
11:24 Calinou    but it's not possible to rotate collision boxes by design
11:24 Calinou    I'd rather see cylinder/sphere collision/selection boxes
11:25 Calinou    both cases are far from trivial
11:28 juhdanad   Then s16. Thanks!
12:41 juhdanad   I made a member variable (core::aabbox3d<s16>) in CNodeDefManager, with a virtual getter in INodeDefManager. This getter is slow, when I repeat it 100 times in every step, my FPS drops from 28 to 14. Can anybody help?
14:03 juhdanad   Can I split getPointedThing() into handlePlayerPointedThing() (updates Hud) and raycast() (actual getPointedThing, but uses Environment and is independent of the client)? Raycast() could be then exposed.
14:03 juhdanad   (exposed to Lua)
14:48 sfan5      juhdanad: you can of course do that but in the end the core devs will decide about code changes
14:49 juhdanad   Yes, I just wanted to know if it might worth trying. Thanks!
17:18 Nosrick    Anyone about that can possibly give me a hand with an error I'm getting?
17:21 juhdanad   What is your error? (I'm not sure that I can help you.)
17:21 Nosrick    I'm trying to load a script using require(), but no matter what I try, it can't find the file I'm trying to reference.
17:23 Krock      Nosrick, http://stackoverflow.com/questions/9145432/load-lua-files-by-relative-path
17:24 Krock      when you specify a name it will automatically search for *.dll *.so and init.lua files
17:24 Krock      maybe you want to try dofile() instead
17:25 Nosrick    Well, I need a reference to the script, so I can access the functions within it.
17:25 Nosrick    And I don't think dofile returns anything
17:26 Krock      you can specify a global table to add your functions there. With dofile() you can extend that table and use the functions wherever you want
17:27 Krock      many mods do that and it works perfectly.
17:27 Nosrick    Got any example of that?
17:28 Nosrick    (As I'm still new to Minetest modding and Lua in general, so that might be a little above my pay grade)
17:29 Krock      Nosrick, check this one: https://github.com/Bremaweb/landrush/
17:29 Krock      in init.lua it creates a new global table, "landrush"
17:30 Krock      and in functions.lua it extends that table
17:30 Krock      (because init.lua uses dofile())
17:31 Krock      oh. another thing. This is the channel for the core development, not for mod development
17:32 Krock      either #minetest or #minetest-project may fit better for this
17:32 Nosrick    Woops, my apologies!
17:32 Nosrick    Thank you so much for your help.
17:32 Krock      no problem. It's too silent in here.
17:41 Hijiri     Nosrick: dofile does return the return value of the script
17:41 Nosrick    Ah, okay!
17:42 Hijiri     If you don't need a global table it is cleaner to use the return value
17:42 Nosrick    I unfortunately do need a global table.
17:42 Krock      that largely depends on your coding style
17:44 Nosrick    Spaghetti.
17:45 Zedicius   Any core devs around? I've might've caught a bug, didn't spot it on the forums and was wondering if it's a known issue or not.
17:45 twoelk     check the issues
17:45 Krock      .. on github ^
17:45 KaadmY     Zedicius: coldn't you just report the bug on the forum or report it as an issue on github?
17:46 twoelk     https://github.com/minetest/minetest/issues
17:46 Zedicius   Didn't find it, though there are over 600 issues.
17:47 * twoelk   wants to know the bug by know
17:47 Zedicius   https://www.youtube.com/watch?v=YUfSUpvTveE
17:48 Zedicius   No mods installed
17:48 KaadmY     items slide forever?
17:48 KaadmY     along edges?
17:48 twoelk     now thats a neat feature :-O
17:49 Zedicius   Depends on the angle. If they go to far off the edge they fall off and stop.
17:49 KaadmY     >.< twoelk, Krock and Zedicius all have purple nicknames on irc for my client
17:49 Zedicius   Is that good or bad?
17:49 KaadmY     a bit bad :P
17:50 KaadmY     hard to see who's who
17:51 Krock      KaadmY, they're all blue here and I have no problem. Try changing your font
17:52 Krock      but change my color to transparent, please
17:52 KaadmY     ;_;
17:52 twoelk     hmm all black here - but then again that's what I set up
17:55 Krock      Isn't that a discrimination? Black latters on white background?
18:00 Zedicius   So! I'll just throw it on the forum and see what'll happen with it  :P
18:04 Krock      Zedicius, could you please check the behaviour with https://github.com/minetest/minetest/pull/4370 ?
18:04 Hijiri     I was thinking of implementing the "attached particle spawners" feature #567, and was wondering what I should do with particle spawners that outlive their objects
18:05 ShadowBot  https://github.com/minetest/minetest/issues/567 -- Implement attached particle spawners.
18:05 Hijiri     ObjRefs currently just become invalid, so I could just remove the spawner and the spawner will be invalid too
18:05 Hijiri     Or I could let the spawner live on, but then if modders want it to be removed at the same time as the obj, they will need to poll for the object
18:06 Hijiri     A more complex solution would be to let modders attach a finalizer that fires when the particle spawner's object dies
18:06 Hijiri     Oh wait, they could do that in the object already
18:06 Krock      to keep it simple for modders, I'd prefer that they would disappear when the ObjRef became invalid
18:07 Hijiri     And let the particle spawner reference become invalid?
18:07 Krock      yes, otherwise it's required by modders to do that manually which is just garbage collecting
18:08 Hijiri     ok
18:09 Hijiri     It seems though that the modder would have to to garbage collection anyway, to get rid of the invalid particle spawner
18:09 Hijiri     assuming they have the reference in a table somewhere, not in an entity
18:09 Hijiri     For example, for players
18:14 Hijiri     I guess letting it go invalid is still better in the sense of less garbage collection though, because you won't need to do it if it's in an object
18:14 Hijiri     I can't think of any use cases where you would actually want it to live on, other than to avoid invalid references
18:15 Zedicius   Krick 4370 still has the same issue.
18:15 Zedicius   Krock  ^
18:16 Krock      okay. going to look into that issue
18:33 Krock      I see. It would increase the node checks from currently 1 to up to 4
18:36 Krock      but it's a cool bartender move :)
18:36 Krock      ^ Zedicius
18:39 Zedicius   Heh. The irony is, I discovered it while messing around with an old conveyor belt mod on the forums...
18:45 Hijiri     Krock: Do all entities that are loaded on the server get sent to clients?
18:45 Krock      only if the player is close to them
18:45 Hijiri     More specifically, can I expect that any entities loaded on the server are also loaded on the client
18:46 Hijiri     ok
18:46 Krock      the setting for that is called "active_object_send_range_blocks" and is by default 3*16 = 48 nodes distance
18:46 Hijiri     So I think there is another issue with automatic invalidation, in the case that an entity with an attached spawner moves close to a player, when it wasn't loaded on the client before
18:47 Hijiri     The client won't know whether an entity was removed or if the entity just got unloaded
18:48 Hijiri     so it won't know whether to unload the spawner
18:48 Hijiri     I mean remove
18:50 Hijiri     unless in the protocol they send different messages for entity removal and unloading?
18:51 * Krock    doesn't know the answer
18:58 paramat    fix for missing jungltrees near chunk borders #4402 see #4401
18:58 ShadowBot  https://github.com/minetest/minetest/issues/4402 -- Decorations: Fix decoration height check errors by paramat
18:58 ShadowBot  https://github.com/minetest/minetest/issues/4401 -- Tall schematic decorations unnecessarily failing near chunk top
19:13 Hijiri     Krock: actually I think this isn't an issue if the server just keeps track of the spawners on each object and uses the ordinary spawner deletion on them when the object is unloaded
19:19 twoelk     minetest is a universe where if a minecart passes by and nobody is there to see it it doesn't exist
19:19 twoelk     hm, rather it pends to exist
19:21 Fixer      wat?
19:21 * twoelk   starts studying a philosophy book about phenology to get this right
19:22 twoelk     argh, wheres my dictionary
19:23 twoelk     Phenomenology?
20:52 Tesseract  I finally managed to get my (Android) immersive mode bug figured out.  I'll be pushing it to my Android update branch in just a moment.
20:53 Tesseract  Turns out I just had to set the system UI flags earlier, before the native app had a chance to get the screen dimensions.
21:17 Fixer      https://i.imgur.com/FMIyC3r.png after lava cast
21:17 Tesseract  est31: Nice job with the X11 icon.  But there's a little issue with it...  http://sprunge.us/hPOU
21:17 Krock      WOHOO! ShadowNinja is back!
21:18 est31      welcome back Tesseract
21:18 Tesseract  Heh, hi :-)
21:20 Tesseract  est31: If you try to use minetest.totallynotanimageformat as the icon it should spin in an infinite loop or crash, depending on what getImageLoader does.
21:21 est31      yeah, I'll fix that
21:22 est31      https://github.com/est31/minetest/commit/e0204c59883c3ac45252262837f57234c8116f7e
21:22 est31      Tesseract, what about this
21:23 Tesseract  est31: That looks good.  Nitpick though: PROJECT_NAME "-icon-128.png"?  It isn't really xorg-specific.
21:24 shamoanjac does anyone know where the code for interpreting texture strings is? I am trying to find it but I get lost somewhere when draw2DImage() is called
21:24 est31      Tesseract, the format may be not, but it is loaded just for xorg
21:24 shamoanjac I have updated issue 4396 about textures in formspecs, apparently it goes further than that, since lower resolutions work
21:24 est31      so its xorg specific, no?
21:26 Tesseract  est31: Well, it's just used for xorg, but the file itself isn't xorg-specific, you can use it for whatever you want.  It hardly matters though, you can leave it if you want.
21:29 shamoanjac ah found it it was tiles.cpp
21:30 paramat    Tesseract rubenwardy has been invited to join the mtgame team, hope you're okay with that
21:30 est31      There is a related bug to it, #4323. Any ideas Tesseract how to solve it?
21:30 ShadowBot  https://github.com/minetest/minetest/issues/4323 -- X11 Icon not shown when minetest is installed via make install
21:30 Tesseract  paramat: Yep, sounds good.
21:31 est31      don't you have to be a dev to be in mtgame team?
21:31 paramat    core dev yes
21:31 paramat    looks like c55 has already made him a core dev
21:31 est31      he's made him a trusted non-dev
21:32 est31      so he's a dev now, good to know.
21:33 Tesseract  est31: You could just add it to line 177 of CMakeLists.txt
21:36 Tesseract  Then pass ICONDIR to the executable in cmake_config.h.  (It's hardcoded to $prefix/share/icons on UNIX)
21:36 paramat    now just waiting for rubenwardy to confirm his acceptance
21:36 est31      He has confirmed it
21:36 est31      https://github.com/orgs/minetest/teams/team-minetest
21:37 paramat    ok i guess he had to accept the invite to that team
21:45 est31      Tesseract, is ICONDIR a good place for it?
21:46 Tesseract  est31: I think so.  The .svg already goes there, so it seems like a good place.
21:47 Tesseract  You'll just have to use the "128x128" directory instead of "scalable".
21:57 est31      Thanks
22:06 est31      Why don't we just say that rubenwardy is an official dev
22:07 est31      i mean "mtgame member but no coredev" that would be an even newer category.
22:09 paramat    he is a core dev now, just hasn't been removed from 'trusted non-core-devs' team
22:09 est31      but you only asked mtgame members
22:10 est31      never asked the other coredevs
22:10 paramat    hm good point
22:10 est31      its not that I would be against rubenwardy to be either coredev or mtgame member
22:10 est31      in fact I support him being a member
22:10 paramat    maybe he should only push to game then, for now
22:11 est31      I just feel a sad that I wasnt asked :)
22:11 est31      a bit*
22:12 paramat    yeah sorry, i guess i was in a way suggesting a mtgame-only core dev
22:13 paramat    we should ask all core devs then, as a core dev has push access to all repos
22:13 est31      and a core dev has +1 "access"
22:15 est31      idk, I don't know what rubenwardy wants
22:15 est31      I dont want to ask all coredevs for a vote until I know what his wishes / intentions are
22:16 paramat    i'll discuss this if i see him around
22:18 est31      maybe I am too buerocratic, but I like things to be clear :)
22:22 paramat    it's a good point
22:29 est31      Everybody PTAL https://github.com/est31/minetest/commit/4ec667190967591727294ebbca0a38259fc006c7
22:29 est31      if no +1, I'll push in 10 minutes
22:30 est31      s/+1/reaction/
22:41 est31      done
22:43 paramat    ah that, +1
22:51 Hijiri     Are the ids of corresponding CAOs and SAOs equal?
23:17 Hijiri     Where is the protocol defined? How would I start extending the particlespawner-adding message to optionally include an object it's attached to?
23:24 Hijiri     nevermind, I think I found it on the dev wikiw