Minetest logo

IRC log for #minetest-dev, 2015-09-25

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

All times shown according to UTC.

Time Nick Message
00:06 waressearcher2 turtleman_: any suggestions of a new name ?
00:08 turtleman_ what kind of tree gives you apples 1/8 of the time
00:08 turtleman_ maybe just call it an apple tree?
00:08 turtleman_ thats the only thing that makes sense to me
00:15 waressearcher2 turtleman_: it can still be that "minecraft" influence so for time being it will be called "tree"
00:16 waressearcher2 turtleman_: like you don't call "dirt" block a "soil" or "earth"
00:26 turtleman_ huh? im a little confused
00:26 turtleman_ im just saying that there are specific trees now, like pine, acacia, and jungle
00:26 turtleman_ i just want things to be symmetric.
00:31 T4im you do have a point with the name; perhaps suggest it in the minetest_game issue tracker? that way it will be seen by more people
00:33 turtleman_ ok
00:33 T4im https://github.com/minetest/minetest_game/issues
00:33 T4im :)
00:53 turtleman_ https://github.com/minetes​t/minetest_game/issues/684
00:58 paramat joined #minetest-dev
01:02 paramat although nice to have specific tree types, changing the code is too disruptive. tree is some kind of appletree. the best match for jungletree seems to be rosewood
01:10 turtleman_ paramat, what do you mean? are you opposed to the new trees?
01:11 paramat i like the idea of specifying the trees, i have thought about that too. changing the code causes too many problems though
01:12 paramat the best match for the dark junglewood colour is wenge
01:14 H-H-H joined #minetest-dev
01:16 paramat we can still possibly consider appletree and jungletree to be specific species, just not change the names in the code =)
01:25 werwerwer joined #minetest-dev
01:38 paramat left #minetest-dev
01:42 turtleman_ jungletree is already in code though, isnt it>?
01:42 turtleman_ i thought they were already separate
01:42 turtleman_ jungle, acacia, pine
01:42 turtleman_ are they parts of the same base?
01:47 waressearcher2 what about bamboo ?
02:10 turtleman_ bamboo sounds cool
02:10 leat joined #minetest-dev
03:10 eugd joined #minetest-dev
03:14 waressearcher2 when someone trying to fix source and upload to main git repository, how can he be sure that no one edited that file allready ?
03:15 asl97 it should error when you push if someone edited that file
03:17 waressearcher2 so in that case you redownload and reedit and reupload it ? what if it gives an error again ?
03:18 asl97 you redownload and reedit and reupload it again xD
03:18 asl97 until it get in,  more or less
03:19 eugd joined #minetest-dev
03:20 waressearcher2 can't you just block the file while you editing and uploading ? so everyone else will wait ?
03:23 asl97 i doubt this kind of thing happen much
03:32 asl97 waressearcher2: as when it error, you just need to git pull it and it should auto-merging whatever edit the other people did (most of the time)
03:54 hmmmm aghh
03:55 hmmmm what kind of suffix should be used to mean "a more advanced version of the same function"
03:55 hmmmm i would use Ex but the meaning of that is muddled thanks to the NoEx suffix
03:57 ShadowNinja Plus?
03:58 ShadowNinja You'll need to document what "more advanced" means though.
04:00 est31 I guess more options or so...
04:04 waressearcher2 function, function_, function_2, function_50
04:05 asl97 waressearcher2: that's postfix
04:05 waressearcher2 use postfix, but sendmail is better
04:08 VanessaE hmmmm: _extended ?
04:08 hmmmm meh
04:08 VanessaE (or just _advanced)
04:08 est31 _ is bad
04:09 est31 we have CamelCase
04:09 VanessaE ok camelCase then
04:09 est31 well, camelCase
04:09 hmmmm by the way about that
04:09 VanessaE either way, don't go digging out the thesaurus  :P
04:09 hmmmm do you guys like our current setup of having ClassName::camelCaseMethod()?
04:09 est31 yes
04:09 hmmmm or would you prefer ClassName::PascalCaseMethod()
04:10 hmmmm I noticed some functions in RemoteClient use the latter
04:10 est31 meh, no, camel better than pascal
04:10 est31 but honestly not that important
04:11 hmmmm of course not
04:12 hmmmm i just want to make sure the code style actually reflects what the developers want
04:12 hmmmm check up every couple of months
04:13 est31 I dont think we should change code style
04:13 est31 bc that leads to first doing it in style a then changing to b then back to a etc
04:14 paramat joined #minetest-dev
04:15 paramat i prefer camelcase, one less shift press
04:35 * hmmmm blinks
04:35 Miner_48er joined #minetest-dev
04:36 hmmmm http://fpaste.org/271312/43155755/
04:38 est31 wtf
04:39 est31 its also funny
04:39 est31 if (pos_ok)  do_a;
04:39 hmmmm // really want to make sure we got the correct node
04:39 est31 if (!pos_ok) { do_b; }
04:40 est31 instead of doing else
04:41 OldCoder joined #minetest-dev
04:41 est31 hmmmm, btw seen what Robert_Zenz has pointed out in your commit
04:42 hmmmm ?
04:42 est31 the param does more than queue control after all
04:42 est31 it sets flags for the queue entry
04:42 hmmmm yeah I know
04:43 est31 okay, you are aware thats good
04:44 hmmmm hah!  i was right
04:45 hmmmm that was zeno
04:45 hmmmm https://github.com/minetest/minetest/commit​/5b8855e83c0d1cc7aef21492e7fe862b7d06917e#d​iff-ad60d65b34e16a3319296bb5d683acd6R2427
04:45 hmmmm no clue how I didn't catch that when originally reviewing this
04:46 hmmmm in any case
04:46 hmmmm I don't see any actual reason why the block needs to be queued into the emerge queue
04:47 hmmmm it's fruitless - if the user is close enough to a node to interact with it, and it has not been loaded yet, you can be certain it'll get loaded soon enough
04:47 hmmmm either that or they're cheating
04:48 est31 yeah, it should be better just be a warning and done
04:48 est31 the result isnt used anywhere
04:49 hmmmm alternatively, it could be a way to get blocks unstuck that somehow get stuck
04:49 hmmmm i wonder if that had been the original intention
04:53 est31 thats fighting the symptoms
04:54 est31 better find out why they get stuck and then fix that
04:54 est31 if they do
04:54 hmmmm of course
05:17 paramat left #minetest-dev
05:48 Darcidride joined #minetest-dev
06:13 nrzkt joined #minetest-dev
06:20 Hunterz joined #minetest-dev
06:29 Player_2 joined #minetest-dev
06:48 VargaD joined #minetest-dev
07:00 VanessaE joined #minetest-dev
07:13 Krock joined #minetest-dev
07:53 julienrat joined #minetest-dev
07:53 julienrat left #minetest-dev
08:32 Amaz joined #minetest-dev
09:12 T4im joined #minetest-dev
09:31 lisacvuk joined #minetest-dev
10:12 Krock joined #minetest-dev
11:00 proller joined #minetest-dev
11:04 turtleman_ joined #minetest-dev
11:06 kilbith joined #minetest-dev
11:18 H-H-H joined #minetest-dev
11:44 Calinou joined #minetest-dev
12:21 lisacvuk joined #minetest-dev
12:46 Calinou joined #minetest-dev
12:47 Amaz joined #minetest-dev
13:18 Player_2 joined #minetest-dev
13:31 Zeitgeist_ joined #minetest-dev
14:03 kilbith joined #minetest-dev
14:05 Zeitgeist_ joined #minetest-dev
14:17 hmmmm joined #minetest-dev
15:12 eugd joined #minetest-dev
15:14 EvergreenTree joined #minetest-dev
15:33 Miner_48er joined #minetest-dev
15:40 rubenwardy joined #minetest-dev
15:48 CraigyDavi joined #minetest-dev
16:14 proller joined #minetest-dev
16:17 eugd https://github.com/minetest/minetest/pull/3199 what's up with the check results here?
16:19 hmmmm check results?
16:19 hmmmm it looks like they passed...
16:19 hmmmm oh I see what you're talking about
16:20 hmmmm yeah I have no idea what you're doing wrong, but it wouldn't hurt to squash all those commits into one and then see what happens
16:22 Soni joined #minetest-dev
16:22 eugd how do i do that?
16:23 hmmmm git rebase -i 9ce30ab
16:23 hmmmm squash, pick pick pick pick ..., rename the commit
16:26 Krock pick, pick, pick, like a chicken
16:28 rubenwardy isn't it    pick squash squash squash squash?
16:28 rubenwardy as pick keeps the commit
16:28 rubenwardy and squash merges into parent
16:28 hmmmm heh maybe it is
16:28 rubenwardy protip: use 's' instead of 'squash'
16:28 hmmmm i haven't found myself needing to squash commits very often
16:29 hmmmm if you do mix them up git would tell you about it though
16:48 proller joined #minetest-dev
17:03 waressearcher2 left #minetest-dev
17:04 Puma_rc joined #minetest-dev
17:10 Gael-de-Sailly joined #minetest-dev
17:12 nrzkt joined #minetest-dev
17:21 Player_2 joined #minetest-dev
17:46 kilbith chuis content, de plus en plus de monde remplace homedecor par xdecor :)
17:47 kilbith fini les maisons de barbie haute définition
17:47 kilbith et quasi toutes identiques
17:48 Calinou je suis très content que tu parles français dans ce canal
17:48 kilbith hein ?
17:48 Calinou tu es sur le mauvais canal :)
17:49 kilbith ah merde !
17:52 proller joined #minetest-dev
17:53 eugd joined #minetest-dev
18:01 proller joined #minetest-dev
18:07 eugd how do sector coordinates work?
18:07 eugd how do they align to block/node coords?
18:08 T4im factor 16
18:08 T4im that is mapblock to node
18:08 eugd yeah
18:08 eugd but when handling sectors
18:08 eugd which block or node is 'center'?
18:10 eugd code is already enabled to tell in debug mode, just don't have machine to do it:(
18:10 T4im not quite sure what you mean with sector, mapchunks? or mapblocks?
18:10 hmmmm erm
18:10 hmmmm 0?
18:10 hmmmm a sector is just a vertical column of mapblocks
18:10 hmmmm block coordinate y = 0 would be the center
18:14 eugd sectors are 80x80x80 but are referred to with single v2s16
18:15 lisacvuk joined #minetest-dev
18:15 T4im 80³ were mapchunks (=5³ mapblocks)
18:15 T4im iirc
18:16 eugd yes you're right
18:16 eugd still 80x80
18:17 hmmmm map chunks are a mapgen abstraction only
18:17 hmmmm there is no "MapChunk" structure
18:17 sfan5 ooh
18:17 eugd yes
18:17 hmmmm it's adjustable via config as well by the way
18:17 sfan5 so thats why i didn't quite get those
18:17 hmmmm this seems to be a common misconception...
18:17 eugd i though what i was asking was straightforward but now i'm confused, ha
18:17 eugd i'm looking at ServerMap::createSector(v2s16 p2d)
18:17 hmmmm there's nothing confusing about this
18:18 hmmmm there are three containers in a heirarchy
18:18 eugd yes sector/block/node
18:18 eugd sector is vertical slice
18:18 hmmmm a Map contains a 2d grid of MapSectors, which are vertical columns of MapBlocks
18:18 hmmmm a MapBlock is a voxel area containing 16x16x16 MapNodes
18:18 hmmmm a MapNode is the smallest unit
18:19 eugd that v2s16 p2d, it is a node address, right? or do sectors start from 1, 2, 3, etc. in their own coordinate system?
18:19 eugd i know the answer
18:19 eugd ugh
18:19 hmmmm be careful here
18:19 hmmmm there are block coordinates and node coordinates
18:19 eugd are there?
18:19 hmmmm block coordinates are what's used to look up MapSectors or whatever
18:20 eugd i thought there was only one coord system?
18:20 hmmmm or address a mapblock inside of the map database
18:21 hmmmm to get the block coordinates from a node coordinate, use getNodeBlockPos()
18:21 eugd ah
18:21 eugd i remember that now
18:23 proller joined #minetest-dev
18:25 hmmmm getSector, getBlock, getBlockNoCreateNoEx, emergeBlock, loadBlock, etc.  anything that operates on a mapblock uses block coordinates
18:25 eugd so there are not also sector coordinates
18:26 hmmmm to make this clear, the parameters to these functions are usually "blockpos" or "bp" or something to denote this
18:26 hmmmm there are no sector coordinates, no
18:27 hmmmm a blockpos v3s16 of, say, (4, 1, -3) would translate to the v2s16 of (4, -3) for the sector position, and then y=1 for the block within that sector
18:51 eugd joined #minetest-dev
18:52 Krock2 joined #minetest-dev
19:10 Amaz joined #minetest-dev
19:19 cvtsx joined #minetest-dev
19:22 asl97 joined #minetest-dev
19:33 lisacvuk_ joined #minetest-dev
19:40 MinetestForFun joined #minetest-dev
19:54 DFeniks joined #minetest-dev
20:10 eugd i've been looking all over code and can't find where sector size is defined
20:12 T4im hmmm defined it earlier as a vertical column of mapblocks, which means x and z are 16 each, and "y" probably spans the entire map
20:14 eugd i mean where in the code the 80x80 node, or rather 5x5 block size of each sector is defined (and more importantly, implemented for mapgen purposes)
20:14 eugd nvm just found it probably lol
20:15 T4im you have a minetest.conf option "chunksize"
20:15 T4im with default=5
20:15 eugd ah
20:15 eugd thanks
20:15 asl97 eugd: https://github.com/minetest/minete​st/blob/master/src/constants.h#L78
20:15 asl97 oh, sector
20:15 T4im and that's for the mapblock size ^^
20:15 eugd it's actually user configurable?
20:15 T4im chunksize, yes
20:16 eugd 'chunk' vs 'sector' being chunks are the cubes used for mapgen, right?
20:16 T4im https://github.com/minetest/minetest/b​lob/master/minetest.conf.example#L691
20:16 eugd and i guess sector just inherits its size from the chunksize setting?
20:17 T4im no, if I understood hmm earlier correctly, then from the mapblock size, that asl97 linked
20:17 T4im due to it being a column of mapblocks
20:17 T4im :P
20:18 eugd oh, wait, waht
20:18 eugd you're probably right
20:18 eugd dunno where this mental block came from, damn
20:22 eugd well found an error
20:22 eugd chunksize not really definable
20:22 eugd or it is and it works fine but generates a constant stream of errors
20:22 eugd 'chunksize not divisible by sidelen, setting sidelen to 16'
20:24 T4im where did you set it?
20:24 T4im in the config? O_o
20:24 eugd yeah it disappeared after close and restart
20:25 T4im maybe the unit is documented wrong.. if it expects the chunksize in nodes, it would have to be a multiple of 16
20:25 T4im though the documentation implies it to be in blocks
20:25 T4im defines it to be*
20:26 T4im close and restart might have actually set chunksize to 16 then ;)
20:26 asl97 https://github.com/minetest/minetes​t/blob/f062bbd7a182233f96c61287d039​7534811627d9/doc/lua_api.txt#L3385
20:26 asl97 it might have been cause of a mod or something
20:27 T4im hm.. yea, but that would then appear in a lua stacktrace, wouldn't it?
20:27 eugd it's coming from mg_decoration
20:27 T4im well then.. :D
20:28 eugd note i'm also using map_generation_limit simultaneously
20:28 eugd probably something to do with it
20:28 eugd have it set up to only generate a single block
20:28 asl97 T4im: why should it error when the doc already said it would just set it to the same as chunksize?
20:29 T4im because there's no warning log level to report on it? ;)
20:29 eugd it's strange also that it's not constantly doing it even at the defualt 5
20:30 eugd 5 is also not evenly divisible by 8
20:34 eugd gtg
20:53 lisacvuk_ joined #minetest-dev
21:01 hmmmm eugd, if you want to limit the map to a single 16x16x16 block, you need to set chunksize = 1 as well in the config
21:01 hmmmm except that's not going to work either because it's expected that there's at least room for the map chunk borders
21:46 proller joined #minetest-dev
21:51 paramat joined #minetest-dev
21:53 VanessaE so, what's the story with #3166 ?
21:53 ShadowBot https://github.com/minetest/minetest/issues/3166 -- Send to clients only changed node metadata instad of whole mapblock by RealBadAngel
21:53 Gael-de-Sailly joined #minetest-dev
21:56 paramat eugd hmmmm that error is my fault, i am using a sidelen of 80 for some decorations. i found this necessary because decoration code fails at very low densities (acacia tree, cacti), the larger sidelen helped. however it is better to alter the decoration code, i'll work on this
21:58 paramat i haven't seen RBA around for a while
22:25 samuellwn_ joined #minetest-dev
22:29 paramat problem is .. to work with all chunksizes (without printing an error) sidelen cannot be larger than 16. at low density deco_count can only be 0 or 1 per 16x16 node area which looks bad
22:30 paramat so i will continue to use sidelen = 80. for low densities the code does the right thing setting sidelen = carea_size, maybe the error should not be printed as it's not really an error
22:32 Kray joined #minetest-dev
22:39 paramat .. yes, i'll open a PR
23:05 Siva joined #minetest-dev
23:06 Siva_Machina joined #minetest-dev
23:45 paramat #3203
23:45 ShadowBot https://github.com/minetest/minetest/issues/3203 -- Decorations: Remove error message 'chunksize not divisable by sidelen' by paramat

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