Minetest logo

IRC log for #minetest-dev, 2017-04-18

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

All times shown according to UTC.

Time Nick Message
00:02 Warr1024 joined #minetest-dev
00:13 Warr1024 joined #minetest-dev
00:19 Hijiri started working on toolcaps override
00:19 Hijiri is it ok if it's a separate method instead of a meta field
00:19 Hijiri since it isn't a string or number or other easily meta-able thing
00:26 Hijiri not that I'm waiting or anything, I'll just get feedback on the PR if someone doesn't like it
00:30 GreenDimond joined #minetest-dev
00:42 Taoki joined #minetest-dev
00:42 Hijiri is there something like luaL_checktype except it's ok if it's nil?
00:48 sofar lua_isnil?
01:07 proller joined #minetest-dev
01:50 Hijiri sofar: I meant something that errors if the argument isn't nil or the provided type
01:50 Hijiri I just ended up using lua_isnoneornil and lua_istable
01:50 Hijiri and then luaL_typerror
02:24 kaeza joined #minetest-dev
02:40 diegom joined #minetest-dev
02:52 kaeza joined #minetest-dev
03:09 VanessaE ok, so..
03:09 VanessaE the mesh cache thing is working reasonably well even for my box
03:10 VanessaE we can haz auto view range next? :)
03:11 * benrob0329 slips a not under the door
03:11 benrob0329 it reads: "World Portals"
03:13 VanessaE really though, if that and the huge delay that comes from loading a block with lots of entities in it could be solved, that would be great. :)
03:14 nore VanessaE: isn't it only displaying the entities that is slow?
03:14 VanessaE I can manage 92 to 96% hit rate on the cache, btw
03:15 VanessaE nore: well having them displayed costs FPS yes, but in this case well...  go on VE-S, /spawn, turn and face east, and run
03:15 VanessaE you'll get a few big 1-2 second stalls in the renderer
03:15 VanessaE I assume this is from entities' meshes being created
03:15 VanessaE because indeed, there are lots of entities in that direction (item frames and signs)
03:20 nore hmmm
03:34 cheapie Is it normal that I get the exact same FPS regardless of the undersampling setting? I thought it was supposed to help.
03:36 octacian joined #minetest-dev
03:46 nore cheapie: I think it helps with drawtime, but if the limit is somewhere else it won't help
03:47 sofar cheapie: what fps are you getting before and after?
03:47 cheapie About 26 and 26.
03:47 sofar do you have vsync enabled?
03:47 cheapie (at the spawn on VE-S)
03:47 cheapie Yes (driver forces it on), but I can get 60 FPS elsewhere on the same server.
03:48 Zeno` cheapie what OS?
03:48 cheapie Debian sid amd64, using an R9 380 with amdgpu.
03:54 sofar cheapie: what value for undersampling?
03:54 cheapie I tried 0, 2, and 4. The visual quality changed as expected, but the frame rate was the same on all three.
03:59 sofar well, that's interesting
04:42 nore pushing https://github.com/Ekdohibs/minetest/commi​t/88f99f32c096b17154548de8bfa53d4e9ddbfcf4 is 10 minutes (trivial bugfix)
04:43 sofar std::string::npos
04:43 sofar otherwise, yes, +1
04:44 nore oops indeed, pushed before compiling :)
04:45 sofar merged #5594
04:45 ShadowBot https://github.com/minetest/minetest/issues/5594 -- Don't make TAB exit game if bound to inventory. by sofar
04:47 nore ok, bugfix pushed as well
04:48 sofar ok what PRs am I reviewing/testing tonight?
04:48 sofar player backend thing I think, again?
04:49 nore sofar: #5373 as well
04:49 ShadowBot https://github.com/minetest/minetest/issues/5373 -- Add the possibility to skip light update and fix light later by juhdanad
04:49 nore already tested, only needs code review
05:05 nerzhul joined #minetest-dev
05:09 Hunterz joined #minetest-dev
05:11 sofar I'd be interested to see how players in sqlite3 holds out
05:11 nerzhul sofar, nice thanks for the review, i just need to catch the exception then :)
05:11 sofar tempted to use it
05:11 nerzhul what do you mean by holds out ?
05:12 sofar it's not been used in a production environment yet
05:13 sofar no idea how it'll work with 60k user accounts
05:13 nerzhul oh i'm using pgsql backend since 2 years on my server :p
05:14 nerzhul sqlite with 60k is like having a 3 months old firefox profile
05:14 nerzhul for servers i suggest postgresql instead of sqlite, sqlite is nice for little servers, huge server should really use postgresql
05:15 nerzhul also the player database is tiny compared to a map database :)
05:16 nerzhul i'm on #5607
05:16 ShadowBot https://github.com/minetest/minetest/issues/5607 -- Segfault when drawing load screen
05:18 sofar mariadb instead of postgres? :P
05:24 nerzhul just code it then :p maybe we can add it in next release to complete the SQL backends
05:25 nerzhul after merging this PR we will have a very nice and refactored database backend permitting to plug easily a new backend or a new type of backend for mt :p
05:25 nerzhul for 5607 it's very strange
05:25 nerzhul texture_src is set in game.init
05:25 nerzhul which is called before createClient
05:26 nerzhul Client::m_tsrc is NULL only if set to NULL in Client constructor
05:29 nerzhul also in backtrace lines doesn't match with our lines
05:35 nerzhul sofar, i launch migration using VanessaE players now :) seems to work well
05:37 nerzhul migraiton okay on 9160 players :)
05:37 nerzhul db takes 24 MB on disk
05:37 nerzhul (sqlite)
05:48 nerzhul merging #5591
05:48 ShadowBot https://github.com/minetest/minetest/issues/5591 -- Reorder TileSpec. by sofar
05:50 benrob0329 joined #minetest-dev
06:06 nerzhul could be nice to see bigfoot here :(
06:09 sofar I looked at a bunch of structs/classes but didn't spot any weird ones with holes like that unfortunately
06:13 nerzhul nice shot for CPU afinity :) it's very rare to look at processor level optimizations
06:13 nerzhul (hello java/go/php/python/etc :p)
06:20 sofar trick I learned when doing kernel drivers
06:21 Calinou we should make sure to document PostgreSQL setup well, else people won't really bother with it :P
06:34 nerzhul Calinou, yeah, i added it to database backends on the documentation and explain how to use it in minetest, but maybe i should document the postgresql server part too a little bit
06:37 Fritigern joined #minetest-dev
06:39 nerzhul i added a note about postgresql shared_buffer for non postgresql inited to enhance its performance
06:40 * VanessaE hopes someone will merge judnahad's soft overlays patch, and fix connected nodeboxes not doing full 6d facedir :)
06:46 sofar why the heck is our curl client asking for "default_stone.png" and not the $hash in the $uri?
06:46 sofar is that still the default?
06:47 Calinou did we document remote media server creation by the way?
06:47 Calinou (collectstatic, for example)
06:48 sofar obviously docs are lacking
06:48 sofar if I need to get a script with wrong default options from gist
06:48 sofar and read weblogs to see it's wrong
06:50 Zeno` Can I ask what the future of leveldb is?
06:51 Zeno` Also.. if player stuff is stored in the db, it *will* be able to be easily purged won't it?
06:52 Zeno` because on a server 99% of players (ok, I'm making that statistic up but it's probably accurate) only join once
06:52 Calinou we could have automatic purging toio
06:52 Zeno` true
06:52 Calinou probably not turned on by default, but eg. the server admin can make it so that players who haven't joined for a month are automatically pruned from the database
06:53 Zeno` hopefully they don't go on holiday :D
06:53 VanessaE on my servers, that's done weekly
06:53 Calinou pfft, who disconnects during holidays ;)
06:53 Zeno` lol
06:53 VanessaE anyone who left a player file behind, but never gained interact -> deleted.
06:53 Calinou it could also be made that when you're a regular (eg. you get some privilege), you are never pruned from the database for inactivity
06:53 Zeno` VanessaE, yeah... so that's what I'm asking. If most player stuff is moved to the database then this still needs to be easy to do
06:54 Zeno` probably a command line option would be good as well. But I'm not really sure of the direction this is going yet
06:55 VanessaE well I imagine it'll be something like SELECT FROM players WHERE age < [some number of hours/days/etc]
06:55 nerzhul Zeno`, we can add a such option in next release if needed
06:55 VanessaE but that wouldn't filter for players who have no interact.
06:55 Zeno` I think it'll be needed
06:55 nerzhul VanessaE, not exact, we need accounts to be migrated to db, it's not the case :)
06:55 VanessaE nerzhul: pseudocode.
06:56 nerzhul auth.txt is not moved i didn't do it in players to db, it needs a separated backend and, as Lua is very linked with current file backend on auth i didn't provide a such pr. On my server i removed the lua binding to bypass this problem and have accounts in db :p
06:57 nerzhul but SELECT name FROM accounts where last_conn < NOW() - 1 month (postgresql) and call PlayerDatabase::RemovePlayer should fit + AuthDatabase::RemovePlayer
06:58 nerzhul we will have it after release when i will send and we merge authdb to database yes :)
06:58 VanessaE and THAT is why I won't migrate my servers from the player-files backend...  I'd never have been able to figure that out
06:58 nerzhul i think you do find players -mtime +60 -delete: p
07:00 VanessaE nope.
07:00 VanessaE I just delete abandoned accounts.
07:00 VanessaE ones where interact was never gained.
07:00 VanessaE once a week, but I don't delete based on age.
07:00 nerzhul note: if you delete from players auth.txt i s not purged :p same problem
07:00 nerzhul but i can add a field: last_modification
07:00 nerzhul i add it as a todo
07:00 nerzhul will add it to player db backend
07:01 nerzhul todo added on #5475, i will do it today it's trivia l:p
07:01 ShadowBot https://github.com/minetest/minetest/issues/5475 -- Player data to Database by nerzhul
07:05 Zeno` sounds reasonable
07:06 nerzhul see you soon :)
07:08 Krock joined #minetest-dev
07:08 Krock joined #minetest-dev
07:16 kilbith joined #minetest-dev
07:23 kaeza joined #minetest-dev
07:41 TC01 joined #minetest-dev
07:43 nrzkt joined #minetest-dev
09:20 TC01 joined #minetest-dev
09:40 proller joined #minetest-dev
09:52 shivajiva joined #minetest-dev
09:54 YuGiOhJCJ joined #minetest-dev
09:56 red-004 #5492 #5554
09:56 ShadowBot https://github.com/minetest/minetest/issues/5492 -- [CSM] Add local particles and particlespawners. by red-001
09:56 ShadowBot https://github.com/minetest/minetest/issues/5554 -- [CSM] Allow enabling and disabling mods. by red-001
09:58 Krock joined #minetest-dev
09:58 Krock joined #minetest-dev
10:18 VanessaE http://pastebin.ubuntu.com/24406701/
10:19 VanessaE already reported, https://github.com/minetest​/minetest_game/issues/1700
10:19 juhdanad joined #minetest-dev
10:23 DS-minetest joined #minetest-dev
10:36 nrzkt joined #minetest-dev
10:38 nrzkt ty for the review sfan5 i will fix that soon. Also Select * is always a bad practice, you should always use a finite list of cols to your database, i respect DBA with this :p
10:38 sfan5 k sure
10:42 lisac joined #minetest-dev
10:57 nrzkt just a question sfan5 are you okay if we make players into sqlite a default for new worlds ?
11:03 Krock 1) try getting the backend from world.mt
11:04 Krock 2) reset to sqlite3 if invalid, since the map backend defaults to sqliet3 anyways
11:04 Krock ^ my two cents to this topic
11:23 proller joined #minetest-dev
11:26 sfan5 nrzkt: not this version yet
11:26 sfan5 Krock: that won't work for old worlds
11:27 Krock how old worlds?
11:28 sfan5 old worlds use the "files" backend but it isn't specified in world.mt
11:28 sfan5 so if you set it to sqlite3 by default it will ignore the existing files
11:28 Krock ah right :/
11:42 sfan5 nrzkt: i actually meant that you should keep the "You are using old player file backend. [...]" and only replace after the "You should switch soon to SQLite3 or PostgreSQL. See [...]"
11:44 Megaf joined #minetest-dev
11:53 DI3HARD139 joined #minetest-dev
11:56 nrzkt sfan5, oh i see :) re-adding it then :) i misunderstand
12:00 nrzkt sfan5: is this okay like this ?
12:04 nrzkt => https://github.com/minetest/minetest/pull/5475/​files#diff-59f5585fbd782ec57b38b0ee73fa885cR385
12:07 sfan5 yes
12:13 Zeno` what's the reasoning behind removing leveldb support?
12:13 nrzkt on android you mean ?
12:13 Zeno` oh, this is for Android only?
12:13 nrzkt because leveldb was not removed from core
12:13 nrzkt yes, it's compilation time disabled on android
12:13 Zeno` ah ok then
12:13 Zeno` now I understand
12:13 nrzkt clients uses sqlite and will never switch to leveldb :p
12:14 nrzkt heh
12:14 Zeno` yes, that is fair enough
12:15 nrzkt sfan5: nice :) i will add the last remaining point when i will solve my production issue (apache seems to have 15k connections and i need to handle more :p)
12:15 nrzkt (2 apache * 15k conn)
12:16 Darcidride joined #minetest-dev
12:17 RichardTheTurd joined #minetest-dev
12:54 Fixer joined #minetest-dev
13:02 kilbith joined #minetest-dev
13:04 Warr1024 joined #minetest-dev
13:07 kilbith_ joined #minetest-dev
13:08 Megaf joined #minetest-dev
13:36 paramat joined #minetest-dev
13:46 halt_ joined #minetest-dev
13:47 Grandolf joined #minetest-dev
13:56 nrzkt Krock: strange crash, i don't understand how it can happen for him and neither in android nor in my client, do you have the problem on HEAD with windows ?
14:13 Krock no, I can start Minetest just fine
14:15 Taoki joined #minetest-dev
14:18 nrzkt joined #minetest-dev
14:23 paramat can anyone approve #5581 ? simple PR
14:23 ShadowBot https://github.com/minetest/minetest/issues/5581 -- Fix Lua crash if opening door with slab or stair by MarkuBu
14:24 Krock I have no clue how that stuff works but the code looks good (style-wise)
14:27 cheapie CSM idea: Since sending chat messages from CSM is controversial, why not just limit it to chat commands? Being able to send commands from CSM is the only real reason I wanted to be able to send messages from it anyway.
14:28 Krock there's already a flood prevention
14:28 Krock if chatcommands aren't counted there yet, they sohuld
14:28 Krock *should
14:31 XtremeHacker joined #minetest-dev
14:31 nrzkt sfan5: i have the sqlit epart with: creation_date and modification_date, i just need to add same thing on pgsql :)
14:32 paramat yes i don't know how 5581 works either, but it's a fix from the person who coded finepos and is tested, so i +1
14:32 nrzkt cheapie: yes, sending chatmessage with client side limitation + the current server side limitation should do the trick
14:32 nrzkt #5581
14:32 ShadowBot https://github.com/minetest/minetest/issues/5581 -- Fix Lua crash if opening door with slab or stair by MarkuBu
14:32 cheapie *shrug* If you want to make it able to send arbitrary chat messages, that's fine by me too :P
14:33 xtremehacker_ joined #minetest-dev
14:39 octacian joined #minetest-dev
14:41 celeron55 there needs to be a way for CSM and server-side mods to communicate without rate limiting though, does that exist yet?
14:41 cheapie I think that's in the CSM roadmap thingy, right?
14:41 cheapie "Mod channels" or whatever?
14:42 nrzkt celeron55: not at this moment, it's in the roadmap with mod channels but it will not be added before release
14:42 nrzkt this release is CSM teasing mode :)
14:42 celeron55 that might leave us with many mods that communicate to the server using chat commands
14:42 celeron55 as people will want to do that no matter what
14:43 celeron55 is it wise?
14:43 cheapie Heh, I'm mostly just after having it do "//lua minetest.get_player_by_name("cheap​ie"):override_day_night_ratio(1)" for me :P
14:43 nrzkt in fact the mod channel should works with channel between clients, and channel with server mods
14:44 nrzkt a mod (server or client) will subscribe a channel and communication will be multicast to other listeners
14:45 celeron55 so we're accepting all CSM mods wanting to do that in the next release to just use chat commands
14:45 celeron55 right?
14:46 celeron55 they will do it, it's just a question of what they will use to do it
14:49 nrzkt chat commands are server side or client side, a chat command can use a channel to fetch some informations, but it should be async :)
14:50 nrzkt (a CSM chat command)
14:50 juhdanad joined #minetest-dev
14:51 nrzkt imagine: local mod_channel = minetest.subscribe_mod_channel("test")  ; mod_channel:send({"hallo": "hi"}); while not mod_channel:empty(); do local msg = mod_channel:recv() print(msg) done
14:51 nrzkt (yeah it's shellua :D)
14:53 paramat ok i'll merge #5581 #5606 in 2-3 hours
14:53 ShadowBot https://github.com/minetest/minetest/issues/5581 -- Fix Lua crash if opening door with slab or stair by MarkuBu
14:53 ShadowBot https://github.com/minetest/minetest/issues/5606 -- Conf.example: Move some lines to minetest.conf.example.extra by paramat
14:54 nrzkt sfan5: i added creation_date and modification_date on sqlite and pg for player databases :) (i just need to test postgres at home this evening, i don't have a minetest pgsql on my machine atm)
14:58 nrzkt i squashed a little bit #5475, it's available for a last review :)
14:58 ShadowBot https://github.com/minetest/minetest/issues/5475 -- Player data to Database by nerzhul
15:08 red-004 nrzkt, do new worlds use the sqlite by default?
15:10 nrzkt who broke minetest linter
15:11 nrzkt red-004: it's not the case, i can do it if sfan5 and sofar agree
15:12 nrzkt it's celeron55 heh, i will push a commit to add src/mesh_generator_thread.cpp as it's not a correct style atm and will fix src/mesh_generator_thread.h
15:36 YuGiOhJCJ joined #minetest-dev
16:03 proller joined #minetest-dev
16:10 DS-minetest joined #minetest-dev
16:12 cx384 joined #minetest-dev
16:19 Hunterz joined #minetest-dev
16:25 est31 joined #minetest-dev
16:33 Hijiri Is it ok if I remove the protocol argument from ToolCapabilities::serialize?
16:33 Hijiri since there's only one protocol implemented in there anyway
16:34 Hijiri The issue is I'm trying to serialize tool capabilities to put in item metadata, and there isn't necessarily a player to put in
16:34 sofar If it's not in the way, I would leave it in place
16:35 Hijiri It's in the way because I don't have a protocol version to pass in
16:35 Hijiri unless I just pass in latest version or something
16:38 Hijiri otherwise I need to give itemstack meta serialize and itemstack serialize a version argument
16:41 sofar why can't you pass in the latest version?
16:42 Hijiri lost connection
16:42 Hijiri sorry, just caught up with the public logs
16:43 Hijiri sofar: I don't see the point of even having the version argument if I just always pass in the latest version, except not changing other parts of the code
16:44 Hijiri And if the serialization is allowed to be different based on player protocol version then I might misserialize the toolcaps if protocol changes in the future
16:44 Hijiri since it would appear safe for someone to add version-dependent behavior to the method
16:47 sofar have the serialize code by default pick the latest version?
16:47 sofar how costly would it be to add it back later?
16:47 sofar how costly is it to remove it entirely?
16:48 Hijiri I'm trying to find where it's used currently
16:48 Hijiri probably just in item definition send
16:50 Hijiri sofar: by serialize code do you mean item meta or toolcaps?
16:51 sofar I don't know, I'm just asking generic questions about removing a protocol number :)
16:51 Hijiri well, I could instead add protocol version arguments to the functions I'm modifying that I'm trying to add toolcaps to
16:52 Hijiri I feel like that would involve more changes since it's used in more places
16:52 Hijiri I'll just do something and PR it I guess, I'll get more feedback there
16:53 Hijiri There is some other thing that is semi-blocking me though
16:53 Hijiri I'm not sure how I should add something to the serialized item metadata
16:53 Hijiri since the deserialize uses the Strfnd thing
16:56 Hijiri are metadata keys/values already pre-escaped so they won't contain the delimiters used in serialize/deserialze of ItemStackMetadata?
17:03 octacian_ joined #minetest-dev
17:06 nerzhul joined #minetest-dev
17:27 Hijiri sofar: Is there a convention for naming of internal meta fields?
17:27 sofar none
17:27 Hijiri I might just put the serialized stuff in a special-use meta field instead of making it a separate part of the metadata
17:29 Krock well, you better shouldn't use \0x01 to \x03
17:29 Hijiri sofar: alright
17:29 Hijiri Krock: oh yeah
17:30 Hijiri could be a problem since tool caps serialize numbers
17:30 Hijiri I could json string escape it
17:49 nore a special meta field is the best thing, and json-serializing sounds reasonable to me
17:49 Megaf joined #minetest-dev
17:52 Warr1024 joined #minetest-dev
17:52 nerzhul sofar, sfan5 i didn't see if you're okay to make sqlite3 default for players for a new world ?
17:52 Hijiri oh, I guess I can bypass the protocol version thing if I write custom json thing
17:53 sfan5 doing that next release might be a good idea
17:53 sofar ^^
17:53 nerzhul okay not this, for android too ?
17:53 nerzhul because android have no choice :p
17:53 sfan5 since android is mostly singleplayer do we have a simpler alternative?
17:53 sofar leave android as is imho, not many users anyway
17:54 nerzhul not many users in our build yes :p
17:54 nerzhul i can let files, no problem :) i just need to test it with sqlite3, i will do it in ~10 min, i'm testing pg atm
17:54 Hijiri do we have a data structure to represent json
17:54 Hijiri in C++
17:55 nerzhul Json::Value
17:55 nerzhul it's jsoncpp
17:55 sfan5 we only have whatever jsoncpp has for decoding
17:55 Hijiri alright
17:55 Hijiri thanks
18:00 nerzhul okay migration to pgsql works properly with timestamps, just need to connect one time with a random player to see timestamp
18:00 nerzhul oh
18:00 nerzhul i should fix scheme
18:02 nerzhul on VanessaE migration failed this time
18:02 nerzhul wtf
18:02 nerzhul position = (-2.14748e+07,-467809,3331.97)
18:02 VanessaE yeah.
18:03 nerzhul it seems pgsql doesn't like it with current precision, i will adjust scheme
18:03 Krock Y position outside limits o.o
18:03 VanessaE there's some bug in the engine that causes players to occasionally end up with bullshit coords
18:03 VanessaE that's what leads to crashes in areas mod and a couple others.
18:03 ssieb joined #minetest-dev
18:05 nerzhul 10,4 => 15,7 in numeric precision on pg solve the problem :)
18:05 nerzhul okay all are migrated, and when you start a migration no problem to continue when you have problem and solve it
18:07 Warr1024 joined #minetest-dev
18:09 nerzhul VanessaE, https://github.com/minetest/minetes​t/pull/5475#issuecomment-294932648 for you :D
18:09 nerzhul delete from players where modification_date < NOW() - '2 month' cascade;
18:09 nerzhul :D
18:11 nerzhul only Android should be tested now
18:13 nerzhul i'm compiling
18:13 nerzhul android compilation is stupid, i forget a file in Android.mk it recompile all MT sources...
18:15 Krock same with MSVC
18:15 nerzhul MSVC have a precompiled no ?
18:16 nerzhul the main problem with android, how can i verify the sqlite db xD
18:16 Krock huh? Talking about Minetest. It recompiles all Minetest source files
18:17 nerzhul okay android works well when i set sqltie3 by default
18:17 nerzhul i stopped and restarted the new world it's okay
18:27 nerzhul sfan5, sofar everything is working as intended :) we can merge when you want, android okay, pgsql okay after fixing VanessaE coordinates into pg scheme :)
18:27 VanessaE nerzhul: um, shouldn't someone fix the bug that led to those out-of-bounds coords?
18:27 nerzhul oh it's sure, but it's not the purpose of this PR :)
18:28 nerzhul i ensure we store all coordinates, right or wrong
18:29 nerzhul thanks sfan5 and sofar for your time to test this backend :)
18:38 Megaf_ joined #minetest-dev
18:40 juhdanad joined #minetest-dev
18:47 sofar joined #minetest-dev
18:51 kilbith joined #minetest-dev
19:04 halt_ joined #minetest-dev
19:05 Grandolf joined #minetest-dev
19:07 ShadowNinja sofar: Does using index.mth fix that?
19:07 ShadowNinja Er, the "load by name instead of hash in URL" thing.
19:11 ShadowNinja (see mt_media_collector in my GitHub if you need a way to generate it)
19:15 VanessaE I pointed him to my variation of sfan5's scripts.
19:15 VanessaE idk what he finally did though (forgot)
19:16 kaeza joined #minetest-dev
19:16 sfan5 people often stumble upon the fact that index.mth needs to be served to POST reuqests
19:17 proller joined #minetest-dev
19:18 VanessaE ...which frankly I don't understand the logic behind that
19:35 nerzhul minetest data server is a little bit complicated for nothing :(
19:40 paramat joined #minetest-dev
19:50 rubenwardy joined #minetest-dev
19:50 cx384 joined #minetest-dev
20:10 sofar ShadowNinja: generating the index.mth doesn't seem to be the problem (I can make one that looks just like VanessaE's)
20:10 sofar somehow the problem seems with the POST method
20:11 sofar I'm writing an actual fcgi daemon right now to do it the *right* way
20:11 sofar well, not right now since we had a power outage and my box at home is offline
20:11 sofar I actually want to make it dynamic and a drop-in thing
20:12 sofar so it can handle and manage the cache at the same time
20:12 ShadowNinja Ah.  I'd probably use PHP to just get it to work.
20:12 sofar I'm writing it in go
20:12 sofar so you can use it without php on a server
20:12 sofar mostly to just stuff it behind nginx easily
20:13 sofar and rely on the nginx code to handle the GETs
20:20 paramat are MTG devs ok with me adding tin ore and more realistic crafting for bronze? (for example crafting 1 copper ingot with 7 tin ingots for 8 bronze or whatever) should be quick and easy for me to do, i'll take stuff from moreores
20:20 VanessaE fuck. yes.
20:21 VanessaE (ok I ain't a mt_game dev but so what :P )
20:21 benrob0329 Ditto
20:23 paramat ^^^ sofar rubenwardy nore sfan5 ShadowNinja
20:23 sfan5 dont care
20:24 proller joined #minetest-dev
20:25 nore paramat: ok for me
20:25 nore also, we really need to discuss about the "split" of mtg some time
20:27 paramat and to be clear, no mithril, for simplicity and to avoid trademark issues
20:28 VanessaE what about silver?
20:28 VanessaE can't see a trademark concern there :)
20:29 paramat maybe later, the bronze issue is higher priority
20:29 VanessaE as for mithril, one could simply rename it to say azurite
20:29 VanessaE but ok
20:29 paramat we don't really need another magical ore
20:30 VanessaE my though is simply that if you add tin, that makes three ores (and four metals) formerly handled by moreores. at that point, you should bring in the other two
20:30 VanessaE thought*
20:31 VanessaE i.e. don't leave moreores as this tiny little stub that ends up needing to be absorbed into some other mod
20:32 paramat that's not strong reasoning though, we should add what's needed instead of worrying about how a mod ends up
20:33 VanessaE we should add what's needed AND what's just kinda nice to have
20:34 Shara How can mithril be trademarked when it's used in many different things?
20:34 VanessaE if we only added what's needed, we'd only have ... well... the minimal game.
20:34 VanessaE Shara: JRR Tolkien came up with it, used it in LotR
20:34 paramat more ores could always have new ores added to it, so no issue
20:35 paramat yeah i have no idea if there are trademark issues
20:35 VanessaE paramat: I really don't understand the resistance here.
20:35 Shara VanessaE: a quick google search implies he probably didn't though
20:35 VanessaE why so much worry over having a few more nodes?
20:35 Shara But either way, it has been used in other things too.
20:35 rubenwardy paramat, more realistic ores etc is cool
20:35 rubenwardy not sure about fantasy ores
20:36 nore http://tvtropes.org/pmwiki/pmwiki.php/Main/Mithril
20:36 VanessaE but that's why I said to rename it to azurite:  https://en.wikipedia.org/wiki/Azurite
20:36 VanessaE non-fictional, non-magical.
20:36 nore The name "mithril"note  or similarly spelled variations (mithral, mythril, and others) is present in other fictional contexts like role-playing games, since the Tolkien Estate did not trademark the term, unlike "Hobbit" or "Balrog".
20:36 VanessaE (didn't know it could be a source of copper - that can be very useful in fact)
20:37 nore so, no copyright issues
20:37 cx384 I think we need new groups: ore, lump, metal, and crystal
20:39 paramat ok
20:41 VanessaE ok?
20:43 paramat MTG just has basic fundamental ores, so silver may be ok, but azurite may be excessive. each ore slows mapgen too. mods are there for more special ores
20:44 paramat ok and thanks on the trademark issue
20:44 Sokomine joined #minetest-dev
20:45 cx384 and you should maybe change the crafting recipe of locked chests so that it needs group metal
20:46 ShadowNinja paramat: I'm fine with tin.
20:47 cx384 because the new ores don't have much uses
20:47 benrob0329 cx384: that would work, but not all metals are suited for lockmaking
20:48 benrob0329 Id like to have an actual "lock" that you can lock doors n such with
20:48 Megaf_ joined #minetest-dev
20:48 cx384 hm you're right
20:49 benrob0329 along with a couple more doors for defferent types of wood
20:49 benrob0329 *different
20:51 paramat ok will work on the tin
20:52 VanessaE paramat: each ore slows down the mapgen - by an imperceptible amount.
20:52 VanessaE so really, you shouldn't be worrying about that.
20:52 VanessaE if you think an ore just doesn't make sense to include, that's fine
20:53 VanessaE but arguments toward performance really don't follow - the mapgen's so fast you could add a hundred new ores and not notice a difference.
21:00 cx384 I think an iron copper alloy is unrealistic. https://www.quora.com/Is-there-an-a​lloy-purely-made-of-iron-and-copper
21:03 cx384 Tumbaga is more realistic. https://en.wikipedia.org/wiki/Tumbaga
21:04 sofar paramat: +1 on tin for bronze
21:04 sofar paramat: but, can I be honest? can we please also fix *steel* to be iron + coal?
21:04 halt_ joined #minetest-dev
21:05 sofar steel = iron + coal
21:05 sofar bronze = tin + copper
21:05 nore what about the future of mtg, first? :)
21:05 sofar I put it on the mtg meeting notes for a topic
21:06 sofar we should do a whole meeting on it
21:06 nore yes
21:07 nore it's easier to do it as well since there are less devs
21:08 cx384 do you want to add an allow furnace or something like that?
21:09 sofar I actually would like to add more "crafting" nodes
21:09 paramat humanly imperceptible or unnoticeable periods of time are irrelevant, for example 10ms is unnoticeable but a long time in code. small amounts of time add up, so yes, an ore has to be justified. it's down to priciple: don't carelessly add lots of unnecessary stuff
21:10 VanessaE "don't carelessly add" is fine
21:10 VanessaE but the whole point of performance is how it looks to the end user
21:11 VanessaE it doesn't matter how fast or slow something is "in code".  it matters how fast or slow it looks to the user
21:11 VanessaE and the mapgen is fast as hell now
21:11 paramat fixing steel could be ok, we can consider that separately
21:11 kaeza joined #minetest-dev
21:12 VanessaE sofar: I could be okay with that coal+iron but you really REALLY need a proper alloy furnace for that
21:12 sofar right
21:13 paramat "it doesn't matter how fast or slow something is "in code"" < nonsense. how it seems to the user is the combination of careful optimisations, each saving unnoticeable amounts of time
21:13 VanessaE yes
21:14 VanessaE but saying that 10ms is "a long time in code" doesn't follow
21:14 paramat yeah a furnace with multiple inputs would be god
21:14 paramat good
21:14 paramat even
21:14 VanessaE is that 10ms per block?  per chunk?  for one ore?  or for a hundred?
21:15 * paramat is godless :]
21:15 VanessaE technic has a fuel-fired alloy furnace that might be suitable for the purpose.
21:15 VanessaE (it is as far as I know, just a modified default furnace)
21:16 VanessaE sorry to harp on this paramat, but what good is high performance if you don't USE it? :)
21:17 sofar I'm a fan of "tiering" content
21:17 sofar so e.g. stone cobble makes a regular furnace
21:17 VanessaE bricks make an alloy furnace in technic I think
21:17 cx384 why shouldn't it be possible to smelt alloys in a normal furnace?
21:17 sofar but you need perhaps granite or some rare rock to make an alloy furnace
21:17 VanessaE cx384: because it only takes one input :P
21:18 sofar alternatively, we could upgrade the furnace with something to an upgraded version of the furnace
21:18 cx384 VanessaE, but you can change it.
21:18 VanessaE sofar: so if you smelt iron ore, will you get a plain iron ingot?  if so, why not combine that with bricks (think of the iron as a cheap rebar) to make an alloy furnace?
21:21 nore also note that the game should not be too linear to be interesting
21:21 VanessaE cx384: it can be changed, but I think having a separate machine for the purpose is a better idea.
21:21 nore i.e. you can develop in different directions if you want, even if to reach the best you can do you have to do all of these
21:22 nore the width should increase then decrease as you advance
21:23 cx384 VanessaE, ok but normally you need a blast furnace to smelt iron https://en.wikipedia.org/wiki/​Iron#Blast_furnace_processing
21:23 AntumDeluge joined #minetest-dev
21:24 paramat 10ms is a long time in code anywhere, obviously. you're making zero sense. we are using the performance. we don't need another set of tools
21:24 VanessaE paramat: who said anything about adding more tools?
21:25 VanessaE and 10ms is only a long tine "in code" if you're doing something that ought be fast and isn't.
21:25 VanessaE remember, I came from the days when we counted microseconds, not milliseconds...
21:25 paramat what would azurite be used for in MTG?
21:26 VanessaE off the top of my head, it could be another copper source, given how valuable that metal is in some situations, or a source of blue dye.
21:26 VanessaE it could just be decorative.
21:26 paramat we already have copper ore
21:26 VanessaE or someone could find some weird but crafty use for it
21:27 VanessaE let's put it this way:
21:27 nore <VanessaE> it could just be decorative. <-- like lapis lazuli in MC :p
21:27 VanessaE why do we have metal blocks?
21:27 VanessaE what use are those in mtg?
21:28 VanessaE nore: now that you mention it, I guess MC uses lapis for those same things.
21:28 nore VanessaE: blue dye, decoration, and magic IIRC
21:28 paramat ores in MTG tend to be the basic fundamental ones, more specialist ones are for mods because many don't want the special ones, that's how MTG works
21:28 VanessaE part of what makes mtg "boring" as I've heard people describe it, is that you just can't get very creative with it.
21:28 nore hmm
21:29 paramat they can add mods or use another subgame
21:29 VanessaE and so they do.
21:30 VanessaE paramat: you didn't answer my question though
21:30 VanessaE why have metal blocks?
21:30 paramat anyone who finds MTG boring is actually facing their own lack of imagination, it's like saying LEGO is boring, it's down to what you do with it
21:30 VanessaE what are THOSE used for in mtg?
21:30 Fixer breaking, mtg is boring
21:31 nerzhul lego is boring if you don't have many pieces :)
21:31 paramat our metal blocks are basic metals used for construction, they're in because they're basic
21:31 VanessaE as for LEGO, it IS boring if all you have are a dozen colors of 4x2 bricks.
21:32 nore exactly
21:32 VanessaE paramat: wait, a gold block is basic, but silver isn't?
21:32 paramat MTG is not quite that bad
21:32 nore whereas when you have gears, axles, etc LEGO are not boring
21:32 paramat i'm not opposed to silver
21:32 cx384 I think sometimes minetest can be boring because it is not very challenging
21:32 paramat just azurite or mithril
21:33 nore same with mtg: add some things that the player has to think to build
21:33 XtremeHacker joined #minetest-dev
21:34 nore so, just adding some game mechanics that will cause the player to think how do design something will make it much more interesting
21:34 paramat mtg has a pretty good variety of elements already
21:34 Fixer paramat: for sandbox
21:34 nore paramat: of *elements* yes, but not of mechanics
21:34 VanessaE paramat: ok, then add tin and silver.
21:34 Fixer mine blocks @ build house @ end of game
21:35 nore there are liquid flooding, carts, and I guess that's all
21:35 nore ah, and tree growing
21:35 VanessaE Calinou: mind submitting a PR to merge mithril into gloopblocks so that moreores can be phased out?
21:35 nore and you can't even interface those together
21:35 paramat silver will be considered separately, i'm neutral on that
21:36 nore but anyway I guess we need that mtg splitting and make it a **real** game
21:36 cx384 it takes 10-30 minutes to get the best tools.
21:36 Fixer okay-okay, you can build a very simple railroad and ride it :)
21:36 VanessaE we tried that split-it-up before, nore.  it did not end well.
21:36 Fixer but look on actual server, nobody in their mind use vanilla
21:36 nore (hint: it's not one.)
21:36 Fixer nore: +1
21:37 nore VanessaE: yeah, there were common mods, these were clearly a mistake
21:37 nore what I meant is more like a fork of mtg
21:37 nore and the fork's objective being primarily singleplayer
21:38 nore with breaking changes being allowed
21:38 nore so it can go forward
21:39 nerzhul just do it after release then, don't forget i embed minetest_game in Android build, and if we should switch how manage this for android users ?
21:39 nore yeah, I guess when we do that, use the new game for the android version
21:40 nore the other version will mainly be for servers anyway I think
21:40 Calinou VanessaE: doesn't More Ores have more ores than just mithril?
21:40 Calinou silver and tin
21:40 VanessaE Calinou: it won't soon
21:40 VanessaE tin is going into mtg
21:40 VanessaE and silver may also
21:40 VanessaE that would leave moreores with only mithril.
21:41 Calinou nice
21:41 Calinou I don't think I can look at this until this weekend though
21:41 nore so moreores should be renamed moreore :p
21:41 VanessaE one_more_ore ;)
21:41 nore yetanotherore
21:58 octacian joined #minetest-dev
22:02 nerzhul ooch ! thanks cppcheck spawn_ltree copy TreeDef at each usage
22:03 nerzhul i will provide a PR fixing various memory copy instead of const ref and merge it tomorrow on the morning
22:03 VanessaE that'll cost a bit won'd it?
22:03 VanessaE won't*
22:03 octacian joined #minetest-dev
22:03 octacian joined #minetest-dev
22:04 nerzhul 6 strings, 3 nodes, 5 int and 2 bool
22:14 nerzhul for performance PR: #5610
22:14 ShadowBot https://github.com/minetest/minetest/issues/5610 -- Fix various variables passed by copy instead of const ref by nerzhul
22:14 nerzhul basicly const ref
22:17 sofar nerzhul: looks fine to me
22:17 sofar I saw that spawn_ltree one, too
22:20 nerzhul :)
22:20 nerzhul maybe we should add cppcheck to travis too at a moment, when everything will be fixed
22:30 Lunatrius joined #minetest-dev
22:32 torgdor joined #minetest-dev
22:36 nerzhul merging #5610
22:36 ShadowBot https://github.com/minetest/minetest/issues/5610 -- Fix various variables passed by copy instead of const ref by nerzhul
22:37 torgdor joined #minetest-dev
22:40 torgdor joined #minetest-dev
22:55 octacian Is modmgr documented anywhere?
22:57 sofar the in-game one
22:59 octacian I'm talking modmgr.lua from builting. AFAIK there is no documentation.
23:00 sofar that's what I thought ... the code is the docs afaicr
23:00 octacian lol That's what I've come up with
23:00 octacian I'm working on a PR to allow installing mods from ZIP right from the main menu
23:01 Fixer octacian: why installing from zip if you can read it from zip
23:02 sofar minetest doesn't support reading zip files atm
23:02 sofar it would be a new library to link against
23:02 octacian Since when?
23:02 octacian You guys really need to keep the documentation updated with the code
23:02 octacian https://github.com/minetest/minetest/​blob/master/doc/menu_lua_api.txt#L73
23:03 sofar https://github.com/minetest/mine​test/blob/master/doc/lua_api.txt -> no mention for "zip" ?
23:03 sofar oh, menu_lua_api has it
23:03 sofar nice, sorry, I never knew
23:03 octacian So MT currently doesn't support reading ZIP files? Or no?
23:03 sofar looks like it does
23:04 octacian Well, then I shall continue :D
23:04 octacian Now, to decide where to put the button...
23:05 Fixer octacian: stooooop, why can't you put zip in /mods and just use it as usual (without any new buttons)
23:05 octacian Fixer: you can... The point is to make it easier for some new players by simply having a button in the mods tab that allows installing mods directly from MT's main menu
23:06 Fixer octacian: you mean you press install, you search from mod zip, and then it installs it in mods?
23:06 octacian Yes
23:08 paramat please no button feature, players are able to put a zip in /mods
23:08 octacian paramat: Yes, but as I just stated, it makes it easier for new players who are unaware of that
23:09 octacian Plus, we have a button to "Uninstall selected mod", it only makes sense to have an "Install mod" button
23:09 paramat auto-extraction from zip might be ok though. will it result in a zip and an extracted mod duplicate file in /mods?
23:09 octacian I still have to test. All of the code is still there from when the modstore was supported, so
23:10 sofar please a button feature, players don't even know where the hell /mods/ is
23:10 paramat all players know how to put a file into /mods, they've managed it so far, this is excessive spoon-feeding
23:10 sofar heck no
23:10 sofar stop that
23:11 octacian paramat: that's mostly true, but it doesn't hurt
23:11 sofar stop trying to educate players into stupid things like copying files to a folder
23:11 octacian And again, it's more consistent
23:11 octacian :rotfl:
23:11 sofar they want to *PLAY*, not learn basic computer skills
23:11 sofar I want a friggen button myself
23:12 paramat so it would load from computer desktop i guess?
23:12 sofar I don't want to wrestle my file manager, the damn program knows exactly where it needs to go
23:12 octacian Correct.
23:12 Calinou <octacian> I'm working on a PR to allow installing mods from ZIP right from the main menu
23:12 Calinou that's a pretty nice idea, I find
23:12 octacian However, I must say, the MT file picker should at some point be improved :rotfl:
23:12 Calinou like, it can open a native file manager (a built-in file manager is an alternative, but it's a lot of work to get it right, and many users won't like it)
23:12 Calinou and you can select a .zip, and should the .zip be properly packaged, it's automatically installed
23:12 Calinou (and you can uninstall it from GUI too)
23:12 octacian Built-in file manager ATM. I'll have to look around later as to native file manager
23:13 Calinou native file manager will likely require a sizeable amount of OS-specific code, and may cause issues in fullscreen, though
23:13 paramat ok it seems reasonable if it loads the zip from desktop
23:14 octacian I was also thinking installing from URL would be really neat
23:14 Calinou can be a bit risky, probably behind a switch
23:14 octacian I'm not sure how stable core.download_file is though
23:15 paramat ok i guess some players have trouble finding /mods, 'all' was unreasonable of me
23:17 paramat i guess the button would show a list of files in desktop, and you select the one which is a mod? there may be multiple zips in desktop, some not mods
23:18 octacian Yes
23:18 octacian Again though, file picker does need to be improved to not show hidden files and to only show specific file type
23:19 paramat sorry for overreaction, i was unsure whether you were trying to install from the internet, which is problematic
23:20 octacian Ah, yes, I understand :D
23:23 paramat i'll also merge #5560 later
23:23 ShadowBot https://github.com/minetest/minetest/issues/5560 -- Fix tool digging speed limit by HybridDog
23:33 TC02 joined #minetest-dev
23:34 Hijiri implemented tool cap overrides \o/
23:34 rubenwardy octacian, Calinou: https://github.com/rubenwardy/NodeBoxEditor/b​lob/master/src/dialogs/FileDialog.cpp#L13-L18
23:34 rubenwardy https://github.com/rubenwardy/NodeBoxEdit​or/blob/master/src/util/tinyfiledialogs.c
23:35 rubenwardy requires a 2000 line wrapper to handle all the different OSes though
23:48 paramat er, so yes loading a mod zip from desktop and auto extraction is a nice idea
23:59 rubenwardy paramat: https://forum.minetest.net/vi​ewtopic.php?p=264563#p264563
23:59 rubenwardy lvm related

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