Minetest logo

IRC log for #minetest-dev, 2020-04-05

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

All times shown according to UTC.

Time Nick Message
00:03 kilbith joined #minetest-dev
00:07 Wuzzy sfan5: so I tested the translation again. so. you're right, weblate is definitely uptodate now (sorry for being annoying). good
00:07 Wuzzy sfan5: but in the po/de/minetest.po file there are now 2 fuzzy strings
00:07 Wuzzy did you probably update them on weblate after you made the commit?
00:08 Wuzzy or do you plan to pull from weblate another time?
00:08 Wuzzy note the two fuzzy strings are already finished on weblate so german is really 100%, just not the .po file
00:34 kilbith joined #minetest-dev
00:46 realz joined #minetest-dev
01:35 nephele_ joined #minetest-dev
04:09 Poikilos_ left #minetest-dev
06:29 QwertyDragon joined #minetest-dev
07:05 absurb joined #minetest-dev
07:16 calcul0n joined #minetest-dev
07:31 xnaas joined #minetest-dev
07:31 xnaas left #minetest-dev
07:37 Krock on weak graphics, disabling stars does gain a few FPS. Also using bitmap fonts improves speed - due to reduced complexity in drawing
07:38 Krock the font part however is only relevant when there's text on the screen
07:40 Krock the stars are always drawn, not just in the night
07:52 twoelk joined #minetest-dev
07:57 erlehmann joined #minetest-dev
08:39 ShadowNinja joined #minetest-dev
09:01 Beton joined #minetest-dev
09:06 sfan5 Wuzzy: I updated themselves myself earlier yesterday and of course will translations be upated from Weblate before release, that's the entire point
09:06 sfan5 s/themselves/them/
09:06 Wuzzy good
09:15 mizux joined #minetest-dev
09:49 calcul0n_ joined #minetest-dev
10:00 Kimapr joined #minetest-dev
10:55 twoelk joined #minetest-dev
11:21 Fixer joined #minetest-dev
11:48 erlehmann joined #minetest-dev
12:23 QwertyDragon joined #minetest-dev
12:29 Calinou rubenwardy: try "undersampling = 4" in minetest.conf
12:30 Calinou this should decrease the GPU load significantly as it will render at a lower resolution
12:30 Calinou and of course, disable everything you can, even simple stuff such as fog
12:30 Calinou sfan5: I suppose we could have a minetest.conf setting for stars then
12:30 Calinou (or always disable them on GLES)
12:31 sfan5 maybe halving their count would already suffice? not a fan of disabling them entirely
12:39 twoelk joined #minetest-dev
12:51 proller joined #minetest-dev
13:01 sfan5 merging #9582
13:01 ShadowBot https://github.com/minetest/minetest/issues/9582 -- Fix hovercolor not working in global tag by pyrollo
13:52 QwertyDragon joined #minetest-dev
13:58 turtleman joined #minetest-dev
13:59 kilbith joined #minetest-dev
14:30 twoelk joined #minetest-dev
14:59 Kimapr_ joined #minetest-dev
15:02 sfan5 rubenwardy, Krock: when would you be here to do the release today?
15:02 rubenwardy yes
15:02 rubenwardy wait
15:02 rubenwardy anytime
15:03 rubenwardy (missed the when)
15:03 rubenwardy any time that is still today in the UK
15:08 Krock here as well
15:19 sfan5 I guess we can do it right now then
15:21 Krock what shall I do?
15:21 DS-minetest joined #minetest-dev
15:23 sfan5 don't think there's anything right now, I'll import the changes from weblate now
15:30 sfan5 ok, done
15:32 Krock so what's left is merging master into stable and tagging it
15:32 rubenwardy I'm just thinking over the credits - making it fair
15:32 Krock oh! right
15:32 rubenwardy and also merging credits
15:32 kilbith joined #minetest-dev
15:33 rubenwardy should the credit heuristic be a flat number of commits? I'm looking over the list, and it looks OK
15:33 rubenwardy there would be no changes for active contributors, and some changes in previous
15:33 rubenwardy hmm
15:34 rubenwardy bigger issue is not crediting translators
15:34 sfan5 game#2633 in this release, yes or no?
15:34 ShadowBot https://github.com/minetest/minetest_game/issues/2633 -- Weather mod: Tune cloud density variation by paramat
15:35 Krock I'd prefer to leave this tuning after the PR
15:35 rubenwardy hmmm
15:35 Krock s/PR/release/
15:35 rubenwardy is the current behaviour broken?
15:35 sfan5 not broken, no
15:39 calcul0n joined #minetest-dev
15:42 sfan5 5.3 it is then
15:45 sfan5 rubenwardy: didn't you say that a flat heuristic resulted in the same people? or is it about ordering?
15:45 rubenwardy flat results in the same people for active contributors, but different people for previous
15:45 sfan5 oh
15:45 rubenwardy for example, BenDeutsch (rendering, sqlite auth, server-side movement checking guy) would be dropped from the list
15:45 rubenwardy and kaeza and some other people added
15:46 sfan5 hm
15:46 sfan5 despire it's problems I prefer the changed lines heuristic
15:46 sfan5 despite*
15:47 rubenwardy ok
15:47 rubenwardy yeah, it's not perfect but on average it's better
15:58 Ruslan1 joined #minetest-dev
16:04 kilbith_ joined #minetest-dev
16:11 kilbith joined #minetest-dev
16:21 Taoki joined #minetest-dev
16:27 sfan5 #9593 is ready then?
16:27 ShadowBot https://github.com/minetest/minetest/issues/9593 -- Update credits for 5.2.0 by rubenwardy
16:27 rubenwardy yes,  although it doesn't credit translators
16:27 sfan5 unfortunate, we can fix this for 5.3
16:28 rubenwardy I can quickly do that, and add it to Active Contributos
16:28 rubenwardy but I do think they should be in their own section
16:28 sfan5 agree with that
16:32 troller joined #minetest-dev
16:36 rubenwardy Do we still support Windows 7, given it's end-of-life?
16:37 rubenwardy same for W8.1, it's only supported for those who pay for extended support
16:39 sfan5 I'd say yes
16:40 rubenwardy there is interest in maintaining support for it I guess
16:40 rubenwardy Krock: are you still on XP? :D
16:41 Krock rubenwardy: switched entirely in 0.5.0-dev
16:41 rubenwardy ok, so you're on Windows Vista now?
16:42 Krock just build that thing and it'll probably still run on XP
16:42 sfan5 we didn't yet agree not to ship minimal in release build, did we?
16:42 Krock that's yet not definitive?
16:42 rubenwardy we did agree to not ship it
16:42 Krock #6987
16:42 rubenwardy the PR I made still has problems
16:42 ShadowBot https://github.com/minetest/minetest/issues/6987 -- Stop shipping Minimal development test in release builds?
16:43 Krock it's definitive when a PR's merged
16:43 sfan5 ok but it's not merged so we're still going to ship it
16:43 sfan5 yea
16:43 Krock exactly
16:44 sfan5 ok, windows build are done so I'll push out the release now
16:46 rubenwardy https://github.com/minetest/minetest_game/pull/2633#issuecomment-609445239
16:46 rubenwardy !title
16:46 ShadowBot Weather mod: Tune cloud density variation by paramat · Pull Request #2633 · minetest/minetest_game · GitHub
16:46 rubenwardy paramat says that that PR should be in 5.3.0?
16:46 rubenwardy -2
16:47 rubenwardy (MTG 5.2 hasn't been tagged yet)
16:47 sfan5 too late now
16:47 rubenwardy ok
16:49 Krock +1/-1 line in the wiki changelog
16:49 Krock shall I merge and tag MTG, or is someone already on it?
16:50 sfan5 just done
16:50 rubenwardy sfan5 just has
16:50 Krock great :)
16:50 Topic for #minetest-dev is now Minetest core development and maintenance. Minetest 5.2.0 released! Chit-chat goes to #minetest. http://irc.minetest.net/minetest-dev/ http://dev.minetest.net/
16:51 rubenwardy web#201
16:51 ShadowBot https://github.com/minetest/minetest.github.io/issues/201 -- Update website for Minetest 5.2.0 by rubenwardy
16:51 sfan5 you can merge that
16:51 Krock the error is just locally, right?
16:51 rubenwardy yeah, travis builds
16:52 Krock I meant your gem setup
16:52 rubenwardy yeah
16:53 rubenwardy ok, I've tooted, tweeted, twatted, and more
16:53 Krock quack
16:53 rubenwardy quacked
16:59 rubenwardy one thing I haven't done is posted
16:59 sfan5 hopefully I can convince the forum to let me post a topic
17:00 rubenwardy I'll leave that to you  😬
17:02 rubenwardy user says  "MT 5.2.0 feels more responsive than 5.1.1. Great job contributors!"
17:02 sfan5 they must've been very quick to test it in the 10 minutes it's been out ;)
17:03 rubenwardy lol
17:04 rubenwardy what's the verb for announcing on IRC
17:04 rubenwardy tooted, tweeted, posted, chatted
17:04 Krock spreaded
17:05 Krock you could call it a NOTICE
17:05 DS-minetest the thing that sfan did in #minetest felt like shouting, with the capital letters
17:09 Wuzzy joined #minetest-dev
17:18 monks left #minetest-dev
17:19 rubenwardy sfan5: found packaging/rules
17:24 sfan5 finally managed to put the right content into the post form, now I just need to retry it twen-
17:24 sfan5 oh
17:24 sfan5 it worked on the first try
17:34 Wuzzy yeah whats up with these forums? something is really wrong o_O
17:35 Wuzzy btw: woohoo 5.2.0 release! congrats 🙂
17:37 Wuzzy Arch Linux package maintainer has been alerted.
18:06 erlehmann_ joined #minetest-dev
18:08 kilbith joined #minetest-dev
18:27 kilbith joined #minetest-dev
18:27 kilbith traditionally new release means new core-devs....
18:29 DS-minetest Now that we are out of feature freeze, I'd like to remember you of some simple feature PRs, eg. #9423
18:29 ShadowBot https://github.com/minetest/minetest/issues/9423 -- GUIFormSpecMenu: Add basic element highlighing debug feature by SmallJoker
18:35 * DS-minetest will also leave this here #9065 dup di dup di du
18:35 ShadowBot https://github.com/minetest/minetest/issues/9065 -- Load media from subfolders by Desour
18:39 sfan5 I think I'll start with merging the PRs that already have approvals tomorrow
18:47 fluxflux joined #minetest-dev
18:58 calcul0n_ joined #minetest-dev
19:12 Wuzzy rubenwardy: plz bump max version on content db to 5.2 thx
19:14 rubenwardy Wuzzy: done
19:41 kilbith joined #minetest-dev
19:50 riceandbeans joined #minetest-dev
19:51 riceandbeans yo, curious what you guys make of this error, I have 16G RAM and 16G swap and 3 terminals open and minetest running and it crashes OOM
19:51 riceandbeans https://pastebin.com/FqqEZcy6
19:51 rubenwardy that's an error with LuaJIT
19:52 riceandbeans Can you please explain in slightly more detail?
19:52 rubenwardy I'm typing
19:52 riceandbeans Are you saying it's not something within Minetest but rather in something else?
19:52 riceandbeans I'll also caveat that I'm a weirdo and playing ( or trying to since it dies in 5 seconds of gameplay ) on DragonFly BSD
19:52 rubenwardy to resolve that, you can either:
19:52 rubenwardy 1. Uninstall mods which use a lot of memory, such as moretrees/biomeslib/mapgeny stuff
19:52 rubenwardy 2. Switch to a build that uses a GC64 edition of LuaJIT, where this bug is fixed
19:52 rubenwardy 3. Recompile without LuaJIT, and use PUC Lua instead
19:53 riceandbeans This is a fresh install, fresh world, no mods at all
19:53 rubenwardy weird
19:53 riceandbeans All settings turned down
19:53 riceandbeans I just installed like 10 minutes ago
19:53 rubenwardy what minetest version?
19:53 riceandbeans And like I said, I run a minimal setup on my workstation, Openbox for my WM, a couple terminals, one running top to track wtf is going on, one launching minetest, one SSH'd to a VPS to talk here...
19:54 riceandbeans eh, one sec
19:54 riceandbeans 5.1.1
19:54 riceandbeans That's what's available at this moment for binary package install
19:54 riceandbeans There may be an update to that in a couple days when the bulk package build from upstream is done
19:55 rubenwardy the out of memory isn't system memory, it's LuaJIT memeory
19:55 riceandbeans Or well...it's done, but it's tagged staging at the moment
19:55 riceandbeans limits says I have unlimited
19:55 riceandbeans I don't know how to tell LuaJIT to calm down
19:56 riceandbeans I was trying to show off to my kid and then it blew up on me and core dumped lol
19:56 rubenwardy does it happen immediate?
19:56 riceandbeans Nope.
19:56 kilbith joined #minetest-dev
19:56 riceandbeans So, I can navigate the loading screen and go through all the various menus, frobbing whatever knobs freely
19:56 riceandbeans I can generate a map and load into it
19:56 riceandbeans If I stand still it seems to be ok for maybe 1-2 minutes
19:57 riceandbeans If I engage any objects it lasts maybe 3-10 seconds before core dump, same reason every time
20:00 riceandbeans Hmm, I'll check dports real quick
20:05 calcul0n__ joined #minetest-dev
20:05 rubenwardy first bug #9599
20:05 ShadowBot https://github.com/minetest/minetest/issues/9599 -- English language missing from language menu
20:20 Wuzzy lol
20:31 calcul0n joined #minetest-dev
20:38 riceandbeans Current dports has 5.1.1 :\
20:39 riceandbeans :( also staged has 5.1.1 so I guess there was no update to it
20:39 riceandbeans Sadness.
20:40 kilbith joined #minetest-dev
21:00 kilbith_ joined #minetest-dev
21:08 kilbith_ hum, did anyone notice this ~100 ms. lag between the actual node placing and the placing sound?
21:08 kilbith_ this is terrible
21:10 sfan5 in singleplayer?
21:10 kilbith_ yes
21:12 sfan5 sound and node placing (assuming prediction is not disabled) are both produced client-side
21:12 sfan5 riceandbeans: if it's not too bad for performance, have you considered switching to plain Lua instead?
21:12 kilbith_ yes but it used to be perfectly in sync
21:16 kilbith_ well maybe it's my bluetooth headphones idk
21:16 sfan5 bluetooth is definitely the first place i'd look for added latency
21:16 kilbith_ especially with the terrible audio managing on Linux
21:22 kilbith joined #minetest-dev
21:27 Lone_Wolf joined #minetest-dev
21:30 Jordach joined #minetest-dev
21:30 riceandbeans sfan5: how would one do that?
21:30 sfan5 compile from source with -DENABLE_LUAJIT=FALSE as cmake flag
21:31 riceandbeans Eh, let me see what I can do
21:32 riceandbeans config options mentioned docs and examples
21:32 riceandbeans maybe it's under minetest and not minetest_game
21:32 riceandbeans AH, there it is
21:32 sfan5 minetest_game is just a bunch of lua files and textures, nothing to compile or configure there
21:33 riceandbeans Ok, reconfigured, going to compile and reinstall, then try again
21:37 riceandbeans I wonder how intense the dragonfly patches are
21:37 riceandbeans Well, finished building and reinstalling....
21:37 riceandbeans time to test...
21:39 riceandbeans Ok, so far I played for like 2 minutes and no crashes
21:46 Fixer joined #minetest-dev
21:48 riceandbeans So uh, how deep is the world because I dug straight down expecting bedrock and/or fallling out of the world but it's been descending for a long time
21:49 sfan5 about 30000 nodes
21:51 riceandbeans How does one fly? I enabled double tap space to turn it on, but what is the actual button to fly?
21:54 Wuzzy /grantme all
21:54 Wuzzy then press K
21:55 Wuzzy you need fly privilege to fly. on singleplayer you can just grant this privilege yourselves with /grantme fly
21:55 Wuzzy on servers you might not have the fly privilege. its common on creative servers, but not guaranteed
21:57 Wuzzy riceandbeans: see also: https://wiki.minetest.net/Differences_from_Minecraft
21:57 riceandbeans I mean this is just screwing around on a local workstation
21:58 sfan5 or better: /grantme fly
21:58 sfan5 using /grantme all will also enable creative mode
21:58 sfan5 and one more thing: #minetest is the preferred channel for user support, this is the development one
22:00 riceandbeans Well...the good news is it seems to be working. The bad news is I can't get the dports guys to change the default compile option, so I have to actually debug the luajit and make a patch
22:01 sfan5 there isn't anything to debug really
22:01 fluxflux joined #minetest-dev
22:01 riceandbeans So, the next question would be is it the minetest usage of luajit on dragonfly, or is it something broken in luajit itself on dragonfly.
22:02 riceandbeans I have nothing else I know that uses it to test
22:02 sfan5 luajit uses the upper 33(?) bits of addresses to store other data, so it can only use memory that is available with an address under 2 GB
22:02 sfan5 but if the OS can't give it that exact memory it runs out .. even if it has only allocated a couple of MB
22:03 sfan5 this isn't specific to Minetest, DragonflyBSD or Unix-like systems
22:03 sfan5 here's an existing issue thread: https://github.com/minetest/minetest/issues/2988
22:13 riceandbeans That's disheartening.
22:14 rubenwardy next top priority: #9066
22:14 ShadowBot https://github.com/minetest/minetest/issues/9066 -- Android: add Android Studio support, completely redone build system and java part by MoNTE48
22:14 sfan5 riceandbeans: if you upgrade luajit to 2.1.0-beta(something) you can enable GC64 and get both great Lua performance and no memory usage cap
22:15 riceandbeans luajit-2.0.5_4                 Just-In-Time Compiler for Lua
22:15 riceandbeans Perhaps 2.1 will come to us soon...
22:16 sfan5 I doubt your bsd will want to package a beta version
22:16 riceandbeans Any idea their release schedule?
22:17 sfan5 quite dead
22:17 sfan5 2.1.0-beta3 is from 2017, that should say enough
22:17 riceandbeans Ugh
22:18 riceandbeans Do I dare ask why there's continued usage of a dead lib?
22:18 riceandbeans I read about performance in the one bug, but still
22:18 rubenwardy I mean, Irrlicht is way worse and is required by Minetest
22:18 rubenwardy Irrlicht is the 3d rendering library, and it's last actual release was 2013
22:18 sfan5 lack of alternative, the performance of plain Lua often just doesn't cut it
22:18 rubenwardy they've had backports since, but they've been working on the next release for 7 years
22:19 sfan5 (but don't let this discourage you from playing with plain lua, it's fine in many cases)
22:19 rubenwardy and yeah, LuaJIT is just that much better than Lua. I've never had a OOM, I don't know why it happens so much for many people
22:19 rubenwardy plain Lua will be fine for most people
22:19 riceandbeans What about a fork of luajit to accomplish what's needed for the dead project?
22:19 sfan5 rubenwardy: 1.8.4 is from 2013
22:19 riceandbeans Maybe the same for irrlicht?
22:20 sfan5 but I guess you mean 1.8.1, which is indeed from 2013
22:20 rubenwardy Irrlicht 1.8 = Nov 2012
22:20 rubenwardy 1.8.4 was 2016
22:21 sfan5 riceandbeans: we do not have the capacity and time to maintain a fork of a luajit, especially not at the complexity it has
22:21 sfan5 for Irrlicht: fortunately there are currently no terrible issues that would prevent us from continuing to use the Irrlicht version from 2016
22:21 fluxflux joined #minetest-dev
22:21 riceandbeans sfan5: Well, I didn't necessarily mean you specifically, but I can't imagine you're the only consumers of a dead project
22:22 sfan5 correct
22:22 sfan5 and in fact some people have had the same idea already: https://github.com/moonjit/moonjit
22:23 sfan5 rubenwardy: eh yes meant 2016
22:23 riceandbeans Ok, so call me crazy, but has anyone ever tried using moonjit instead of luajit for minetest?
22:23 sfan5 I've heard it works
22:24 riceandbeans I've never looked at the code I have no idea the complexity within it.
22:26 riceandbeans Also C++ isn't my strong suite. I know several scripting languages comfortably, Rust, and Go.
22:54 kilbith joined #minetest-dev
22:55 rubenwardy 5.1.0 bug report: https://quey.org/@moongaia/103947686435316073
22:56 rubenwardy I'm asking them to update, disregard

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