Minetest logo

IRC log for #minetest-dev, 2016-03-06

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

All times shown according to UTC.

Time Nick Message
00:14 yang2003 joined #minetest-dev
00:27 Taoki Hey there. Is there any news on RealBadAngel? It's been over a week since his heart attack... someone must know something.
00:29 sofar no news, nothing.
00:29 Taoki ok :(
01:13 paramat joined #minetest-dev
01:25 paramat now merging games 891 799 892 880
01:45 paramat merged
01:46 paramat down to 31 PRs nice
01:50 paramat and now attempting merge of game 794
01:52 paramat finished, merge failed
01:52 sofar that should be easy to fix tho
02:11 srifqi joined #minetest-dev
02:15 paramat yes. docs need improving too
02:25 Etzos joined #minetest-dev
02:30 Player_2 joined #minetest-dev
02:58 srifqi joined #minetest-dev
04:06 Miner_48er joined #minetest-dev
04:08 paramat hmmmm what do you think about reverting 'Allow setting chunksize in core.set_mapgen_params'? https://github.com/minetest/minetest/commi​t/5130dbce7b294c392733e93c05e2099786523342 it's used by few and i feel it should be discouraged. also we could remove documentation of the 'chunksize' setting from conf.example and settingtypes.txt to discourage use
04:09 hmmmm uhh, weren't you the one who asked for that commit?
04:12 paramat not sure, but i do remember something was done to chunksize to fix a bug, investigating..
04:13 hmmmm in any case, i don't think chunksize should be settable from the lua api
04:13 hmmmm i never did
04:13 hmmmm somebody made a PR about how they can read chunksize but not set it, i offered an explaination why, and then you gave the ol' "well we'll be really careful, I PROMISE" so then it got added
04:14 paramat oops
04:14 paramat i'll search for what happened
04:15 hmmmm i don't mind what you do, just be aware that it will break existing worlds if you remove that option now
04:16 paramat https://github.com/minetest/minetest/issues/3222
04:16 paramat indeed we need to support it, but disallowing setting it will not break worlds
04:17 paramat i +1 then changed to neutral on the issue
04:18 paramat idiot moi
04:21 paramat do you think it should remain settable in .conf and advanced settings menu?
04:23 paramat anyway at the least i'll revert setting it in 'set mapgen params'
04:28 hmmmm the commit you linked does more than that
04:29 paramat yes i'll leave the other stuff untouched, i won't use 'revert'
04:29 hmmmm how won't this break worlds?
04:30 paramat erm, because it's stored in map meta?
04:31 hmmmm is it??
04:31 paramat yes, checked
04:31 paramat 'Making it a param and being able to 'get' it was done on my request as a way of getting chunksize in certain situations'
04:32 paramat i think it fixed a bug we were having
04:32 paramat iirc
04:32 hmmmm i don't see where any mapgen settings set through lua get stored in map_meta.txt
04:33 paramat it's also set through default settings and .conf or advanced menu
04:33 hmmmm so?
04:34 paramat but i have a world here that contains 'chunksize = 5' in map meta txt
04:34 hmmmm player uses a mod that explicitly sets chunksize through lua
04:34 hmmmm non-default chunksize is not set in map_meta.txt
04:34 hmmmm you create a patch removing chunksize setting functionality
04:34 hmmmm player's world now uses the map_meta.txt stored chunksize, not the mod's override
04:35 paramat ah
04:35 hmmmm now the player gets large chunks of content_ignore between the generated and non-generated land
04:35 hmmmm how is this case averted?
04:36 paramat perhaps we can't revert then
04:36 hmmmm in addition to breakage, the creator of that PR convinced me that it would be a fine addition for his use case
04:37 hmmmm so i think it should stay
04:37 paramat ok agreed
04:37 paramat thanks
05:13 Miner_48er joined #minetest-dev
05:13 Dragonop_ joined #minetest-dev
05:20 bashterm joined #minetest-dev
05:22 bashterm left #minetest-dev
05:25 bashterm joined #minetest-dev
05:25 bashterm left #minetest-dev
05:33 paramat left #minetest-dev
07:28 Hunterz joined #minetest-dev
07:52 Calinou joined #minetest-dev
09:08 davisonio joined #minetest-dev
09:32 Obani joined #minetest-dev
09:42 srifqi joined #minetest-dev
09:46 Krock joined #minetest-dev
10:19 numZero joined #minetest-dev
10:51 davisonio joined #minetest-dev
10:52 davisonio joined #minetest-dev
12:18 turtleman joined #minetest-dev
12:31 misprint joined #minetest-dev
12:45 davisonio joined #minetest-dev
12:55 iqualfragile joined #minetest-dev
13:12 davisonio joined #minetest-dev
13:15 DFeniks joined #minetest-dev
13:16 davisonio joined #minetest-dev
13:59 Player_2 joined #minetest-dev
14:02 Amaz joined #minetest-dev
14:06 Obani joined #minetest-dev
14:26 Megaf joined #minetest-dev
14:38 Fixer joined #minetest-dev
14:39 ud1 joined #minetest-dev
15:03 davisonio joined #minetest-dev
15:06 kaadmy joined #minetest-dev
15:22 blaze joined #minetest-dev
15:24 ElectronLibre joined #minetest-dev
15:35 cd2 joined #minetest-dev
16:30 twoelk joined #minetest-dev
16:37 numZero joined #minetest-dev
16:43 iqualfragile joined #minetest-dev
16:55 iqualfragile joined #minetest-dev
16:58 iqualfragile joined #minetest-dev
17:13 turtleman joined #minetest-dev
17:37 ElectronLibre joined #minetest-dev
17:52 hmmmm joined #minetest-dev
18:00 Megaf joined #minetest-dev
18:10 kaeza joined #minetest-dev
18:14 STHGOM joined #minetest-dev
18:18 sofar I reopened game#758 with different steel door sounds
18:18 ShadowBot https://github.com/minetes​t/minetest_game/issues/758 -- Add sounds for steel doors. by sofar
18:29 est31 joined #minetest-dev
18:29 est31 I see the commit is from oct 2015
18:29 est31 so its not in 0.4.13 yet
18:29 est31 dev stuff can be broken IMO
18:29 est31 users explicitly have to install a mod
18:30 est31 but if hmmmm thinks it should stay, i have no problem with it staying
18:42 sofar I was thinking that it wouldn't be too hard to make a tool that allows you to delete blocks from map.sqlite
18:42 sofar it would come in handy for several purposes
18:42 sofar e.g. regenning map, pruning blocks for a server map, etc.
18:43 est31 in fact there already exist such tools
18:44 est31 from the just test server creator
18:44 sofar doesn't surprise me
18:45 sofar what's the faceposition cache you're referring to?
18:45 est31 some stuff nrz introduced
18:56 celeron55 we probably should have an official multi-purpose tool for managing worlds at the block level
18:56 celeron55 but it's really up to whether someone can make it and maintain it
18:57 celeron55 one notable thing is that hmmmm wanted to make an abstraction layer for world serialization which could have been used for something like that too, but that never happened
18:58 est31 you mean a library?
18:58 sofar having a worldedit method would be nice, but I'm afraid of allowing in-game code to destroy large areas :)
18:58 est31 i dont know whether that will work, isnt it impossible to do C++ libraries
18:58 est31 and a C library, idk whether thats a good idea
18:59 celeron55 of course alternatively the editor could be in minetest itself instead of being a separate tool
18:59 sfan5 or it could be a library and both
18:59 celeron55 none of that is exactly trivial though
19:00 est31 idk we already generate two separate binaries, minetest and minetestserver
19:00 est31 we can have minetestworldedit or whatever its called as well
19:00 celeron55 well maybe the easiest one could be just a bunch of command line options to command minetest to do some bulk operations
19:00 sofar it should be trivial in python to just write a sqlite manipulation thingamabob
19:00 celeron55 sofar: those exist already
19:00 est31 in fact I've done that already for bugfixing some maps :)
19:01 celeron55 i don't know where to find the most recent one though
19:01 celeron55 also OldCoder has been working on something related to this
19:01 celeron55 i don't know how generic it is, but he has said it can be used to combine worlds and automatically remove unused blocks from worlds
19:02 twoelk would be nice if OldCoder would publish what he used to manipulate the Landrush map when he recreated the server
19:03 twoelk or what he used for redcrab server
19:05 est31 there is this https://github.com/AndrejIT/map_unexplore
19:05 est31 with mt_block_parser.py
19:06 twoelk Such a tool could also include something to extract nodes that cause trouble or crash the server without having to actually start minetest
19:07 est31 yeah
19:07 est31 I have done precisely that already
19:07 est31 for the too much entities in a block bug
19:07 est31 but cant find it right now
19:08 est31 thats it https://github.com/minetest/minetest/issues/2610
19:08 est31 code here https://github.com/est31/minetest/​commits/fix_too_many_entities_bug
19:11 ElectronLibre joined #minetest-dev
19:28 est31 has a very fast method to delete all entities in a world
19:39 ShadowNinja Turns out rewriting formatChatLine was easier than trying to decypher it to find an indentation bug.  The result is probably half as long.
19:40 ShadowNinja #3811 should work properly now.
19:40 ShadowBot https://github.com/minetest/minetest/issues/3811 -- Add configurable timestamp to console messages by ShadowNinja
19:41 Fixer ShadowNinja, thanks, i will try it today
19:43 ShadowNinja Incoming patch: http://ix.io/pkS/diff
19:44 twoelk est31: what if I need to downright replace a certain node worldwide without aliasing it or something similar
19:45 est31 ShadowNinja, +1
19:46 est31 ShadowNinja, you want to +1 my LBM PR #3677
19:46 ShadowBot https://github.com/minetest/minetest/issues/3677 -- Add minetest.register_lbm() to run code on block load only by est31
19:47 est31 ?
19:51 Obani With the new chat, ^ character isn't supported anymore
19:51 Obani Is that normal ?
19:51 Obani afk
19:55 est31 ^ works for me
19:55 Dragonop joined #minetest-dev
19:55 ElectronLibre joined #minetest-dev
19:55 ShadowNinja est31: Which one did you +1?  And not quite yet, I'll have to take another look at your latest version, but I GTG now.
19:56 est31 <ShadowNinja> Incoming patch: http://ix.io/pkS/diff
19:56 ElectronLibre joined #minetest-dev
19:56 ShadowNinja est31: OK, thanks, I'll push that once I get back.
19:58 est31 Obani, you want to use ^ to generate stuff like ĉ ?
20:12 sofar #3820
20:12 ShadowBot https://github.com/minetest/minetest/issues/3820 -- Add consistent monotonous day counter - get_daycount() by sofar
20:17 ElectronLibre joined #minetest-dev
20:17 sofar I wish I had thought of that earlier... I've been attempting to implement a sane calendar in a mod for ages
20:18 ElectronLibre joined #minetest-dev
20:20 Etzos joined #minetest-dev
20:22 est31 afk
20:28 juraj_ joined #minetest-dev
20:32 juraj__ joined #minetest-dev
20:32 est31 sofar, perhaps also a setter for the day count?
20:32 est31 and ShadowNinja will probably like it more if its get_day_count
20:32 sofar get_gametime
20:32 sofar but, sure
20:33 sofar I thought about a setter, and it doesn't make any sense to me to have one
20:33 sofar you can edit env_meta.txt if you really want to change it
20:33 juraj__ left #minetest-dev
20:33 sofar enabling time travel seems weird
20:34 juraj_ joined #minetest-dev
20:34 sofar the patch itself is trivial, renaming it is fine with me. ShadowNinja ?
20:36 davisonio joined #minetest-dev
20:36 juraj_ hi, could i please ask question about minetest api and server performance ?
20:36 sofar you can ask in #minetest as well, juraj_
20:37 juraj_ ok thanks
20:38 est31 there is already time travel for the time of day value
20:38 sofar nonsense, that's just nothing more than taking a quick nap ;)
20:41 est31 monotonic monotonous in german you say "monoton"
20:41 est31 nice and short :)
20:44 celeron55 i don't think there should be a setter
20:44 celeron55 it would just make mods mess up each other and be lazy
20:45 celeron55 or, well, dunno really
20:46 est31 perhaps we can consider adding it if there is an actual use case for it
20:49 celeron55 well i'm sure that lazy mods will just reset the counter instead of keeping track of where it started, and mess up other mods by doing so; what i am not sure about is whether there is a valid use case that would overcome that
20:50 sofar that's why I prefer there to be no setter
20:50 sofar besides
20:50 est31 what about a /day command
20:50 est31 /time shows you the current time
20:50 sofar if mods wanted to modify it, they should keep a shadow value
20:50 sofar good idea
20:51 celeron55 i suspect some mods have a /day command that equals /time 12000
20:51 est31 ah
20:52 celeron55 just saying 8)
20:52 sofar we can see if that's the case
20:52 celeron55 maybe there isn't, actually
20:52 celeron55 not a very common one at least
20:52 est31 i think in minecraft the chat commands arent prefixed by /
20:53 est31 and on many servers there are people who try to do commands
20:53 Wayward_One joined #minetest-dev
20:53 est31 often you hear people saying "day" nothing else
20:53 est31 probably its a valid minecraft command
20:53 sofar it's probably a mod to "make it day"
20:54 Calinou in Minecraft you need a / for commands, pretty sure
20:54 est31 hmm right
20:54 Calinou http://users.kymp.net/feuer/etcomic/013.jpg
20:54 sofar in MC, or at least bukkit etc. mods can intercept chat commands and treat them as commands
20:57 est31 ah
20:57 est31 in minecraft it is possible to do commands without slash
20:57 est31 but only in the server console
20:58 celeron55 in minetest you can do absolutely anything based on anything in the chat 8)
20:59 celeron55 (but the engine encourages /commands so that's what everyone uses)
21:04 twoelk some chat mods filter for swear words - you could see swear words as commands that trigger punishment action :-P
21:05 est31 lol
21:06 est31 its always fun watching people avoid those filters
21:06 est31 F U C K  YOU!
21:07 paramat joined #minetest-dev
21:08 sofar so, /day command is 5 lines of code, might as well add it, right?
21:08 SaKeL joined #minetest-dev
21:09 celeron55 as a read-only command, sure
21:09 est31 register_chatcommand i think works in a way where registering the same command again overrides it right?
21:09 celeron55 i imagine it being handy for developing stuff based on days
21:09 est31 or is it with table.insert
21:09 sofar yes, it just displays "Current day is " .. core.get_day_count()
21:09 celeron55 or, testing
21:09 est31 yeah thats what i asked for too
21:09 est31 was in my mind*
21:10 sofar I've got a pending project to do seasonal day/night length variation
21:10 sofar that'll be one
21:10 sofar the second one is a consistent calendar
21:10 sofar like /date
21:11 SaKeLWorld joined #minetest-dev
21:14 paramat nore sfan5 i will soon work on 2 things: make beds white and not dependant on red dye, because roses can be distant in mgv7. add root/branch nodes for jungletrees that have 6 exterior tiles
21:15 paramat and then later for other trees
21:17 kaadmy Calinou: http://users.kymp.net/~feuer/etcomic/023.jpg
21:24 rubenwardy joined #minetest-dev
21:26 Calinou Minetest needs m_pitch
21:26 Calinou for CTF :)
21:27 est31 all the time irrlicht limitations are cited
21:38 paramat nore sfan5 possible close? any comments? game#436
21:39 ShadowBot https://github.com/minetes​t/minetest_game/issues/436 -- Add footstep sounds for liquids by MirceaKitsune
21:39 sfan5 water footstep sounds?
21:39 sfan5 hm
21:40 paramat idea is ok but implementation and sounds here not good
21:43 nore paramat: well, I never use sound when I play a game, so I'm completely neutral on this
21:45 paramat ok, but it's the implementation and sounds that is the problem here
21:47 nrzkt joined #minetest-dev
21:47 nore paramat: lava sounds aren't a very good idea I think
21:48 nore except if you have very good ones (what should even it be?)
21:48 nore about water sounds, well, if you have good ones, then yes
21:49 paramat yeah i might even attempt water sounds myself sometime, but subtle and suitable for underwater
21:49 paramat and one silent sound to break up the regularity
21:53 sofar there's plenty of good sounds on freesound.org
21:53 sofar some are really nice
21:53 sofar just edit them so they can be mixed up nicely
21:55 sofar e.g. add fade-in/fade-out, tune length for overlap
21:55 paramat yeah that's where i would search
21:56 sofar that place is a goldmine
22:02 Icedream joined #minetest-dev
22:02 est31 lol
22:03 est31 in goldmines the gold concentration is less than 1% of the mined material :)
22:03 est31 much less
22:10 sofar the total amount of gold on earth, mined and unmined, would not even fill a soccer stadium
22:13 Fixer What I like in MC, is that it capitalizes first letters and adds . at the end, simulating literate people %)
22:13 sofar Maybe you can implement an IRC filter that automatically modifies chat messages that way.
22:14 kaadmy Fixer: in chat?
22:14 sofar Make everyone seem litterate ;)
22:14 sofar s/tt/t/
22:14 kaadmy ^ heh
22:14 Fixer kaadmy, in chat
22:15 sofar [ot] I've never seen an MC server do that. Vanilla or Bukkit?
22:15 Fixer what I don't like is some strange swastika like animated symbols in chat that rotate and change colour, wth was that?
22:15 Fixer sofar, probably bukkit
22:24 est31 sofar, about 3800
22:24 est31 you said in the commit description: "Protection mods will likely want to mimic the changes in their relevant code sections"
22:24 sofar yes?
22:25 est31 why so
22:25 est31 you can simply replace minetest.is_protected, no?
22:26 sofar protection mods currently have to override minetest.is_protected()
22:26 est31 yes
22:26 est31 thats the intended use
22:26 sofar so e.g. node_dig() calls minetest.is_protected()
22:26 sofar which is fine, since the access priv check is in node_dig()
22:27 sofar but, it won't catch other interactions, chest privs etc. I think
22:27 est31 so what mimicking is required then
22:27 paramat +1 for new steel door sounds
22:28 sofar insert a check for the protection_bypass check before other checks
22:28 sofar *priv
22:28 sofar paramat: ty, yes, that sound was a great find.
22:28 est31 sofar, but if there is one check already inside minetest.is_protected
22:28 sofar protection mods *override* it, the API tells them to
22:28 est31 then why is it still required to have in node_dig
22:29 est31 yes
22:29 est31 isnt it better to just leave it to protection mods
22:29 sofar yes but the is_protected() API was designed wrongly
22:30 sofar you can't chain multiple protection mods
22:30 sofar since the API doc clearly says "protection mods override this function"
22:30 sofar had it been better designed, than it would have been easy
22:30 sofar e.g. minetest.register_protection(func)
22:31 est31 well & chaining still is possible
22:31 sofar it certainly is
22:31 sofar but would require all protection mods to behave
22:31 est31 just || chaining isnt
22:32 sofar it has to be more sophisticated, yes
22:32 est31 well if you dont install a mod that behaves its your problem
22:32 sofar sure
22:32 sofar it's worth redoing imho
22:32 twoelk joined #minetest-dev
22:33 sofar but it also means redesign of the protection mods callback
22:35 est31 I am open to proposaly
22:35 sofar right now they return "allowed" or "denied", but it should be "I don't care", "explicitly allowed", "explicitly denied" I think
22:35 est31 proposals*
22:35 est31 but first I want to see sth better than currently
22:35 sofar the crux is
22:35 est31 there is nil, true and false
22:35 est31 lua ftw
22:35 sofar yes
22:35 sofar that's what I'm thinking
22:36 sofar say, 2 protection mods
22:36 sofar step 1) protection_bypass? yes: done
22:36 sofar step 2) call _all_ protection mods
22:36 sofar if a mod explicitly denies -> deny
22:36 sofar if a mod explicitly allows -> allow
22:37 sofar nothing ? allow
22:37 sofar if one mod denies, other explicitly allows? crap
22:38 sofar of course, if you 'break' when an explicit allow/deny is found, you avoid the problem
22:38 est31 schrodinger solution :)
22:38 sofar protection mods can reorder themselves, too
22:38 sofar bar/depends.txt: foo?
22:39 est31 eww
22:39 est31 that sounds bad
22:39 sofar arms race
22:39 est31 it should rather be left to the person who deploys the mod
22:39 sofar I'm wondering whether it's realistic concern or not
22:40 sofar e.g. someone grants permission to another person in an area that's also protected with a different protection mod
22:40 est31 the other question to extend for the protection api is about actions
22:40 sofar maybe
22:40 sofar explicit allow trumps denies?
22:40 est31 like: you grant access to chests
22:41 sofar so 'deny' -> no break, try more
22:41 sofar yes, something like taht
22:41 est31 but no access to placing blocks
22:41 sofar I think allow > deny could work
22:41 sofar Order allow, deny
22:41 * sofar has bad dreams about httpd.conf
22:42 est31 lol
22:42 est31 we can make it a setting
22:48 paramat shall we merge #3448 before conflicts arise again? sure it's ok?
22:48 ShadowBot https://github.com/minetest/minetest/issues/3448 -- Faster insertion into table by Rui914
22:49 sofar I'd vote merge
22:53 paramat it seems ok to me
22:54 paramat +1
22:54 paramat ok that's now mergeable
22:55 est31 yes
22:57 Dragonop joined #minetest-dev
22:57 paramat 3800 seems ready too
22:59 sofar est31 is having a party with that one
22:59 est31 party?
22:59 sofar hyperbole
23:01 paramat ah ok
23:09 yang2003 joined #minetest-dev
23:19 paramat ok i'll merge 3448 soon before conflicts arise
23:25 paramat and possibly #3816 once i've reviewed it
23:25 ShadowBot https://github.com/minetest/minetest/issues/3816 -- Settings Tab : Regroup dropdown datas in tables by kilbith
23:30 paramat +1 on that
23:39 paramat now merging those 2
23:39 huang2003 joined #minetest-dev
23:51 paramat merged
23:52 paramat ugh down to 128 PRs feels so good
23:55 Fixer lol
23:55 Fixer too optimistic
23:56 rubenwardy now get 64 ;)
23:57 est31 somebody could look at my PR
23:57 est31 with the LBMs
23:57 est31 ShadowNinja is still gone it seems

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