Minetest logo

IRC log for #minetest, 2016-08-29

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

All times shown according to UTC.

Time Nick Message
00:02 Yst joined #minetest
00:02 i- it worked
00:05 MinetestBot [git] red-001 -> minetest/minetest: Settingtypes.txt: Clarify comments, correct spelling fbe4a92 https://git.io/v6jU8 (2016-08-29T01:00:05+01:00)
00:05 MinetestBot [git] SmallJoker -> minetest/minetest: Builtin: Disallow registering users with the same name 51e13ae https://git.io/v6jU4 (2016-08-29T00:58:39+01:00)
00:19 Void7 joined #minetest
00:21 MinetestBot [git] rubenwardy -> minetest/minetest: Fix background formspec elements from interfering with each other 78ff5c1 https://git.io/v6jTe (2016-08-29T01:15:31+01:00)
00:24 agrecascino joined #minetest
00:54 agrecascino joined #minetest
01:17 Taoki joined #minetest
01:19 agrecascino joined #minetest
01:36 Yst joined #minetest
02:08 Freejack joined #minetest
02:25 SusWombat joined #minetest
02:28 Rate85 joined #minetest
02:29 Rate85 Hello?
02:30 Rate85 Something Is Very Wrong Here, This Place Is Usually Filled With Life...
02:30 Rate85 Why Is Nobody On IRC?
02:33 GunshipPenguin joined #minetest
02:42 Void7 actually at least 130 people are on irc (probably)
02:42 Void7 just not very active
02:47 Rate85 I Wonder Why
03:10 paramat joined #minetest
03:12 paramat http://i.imgur.com/AsKvBfu.png pathv7 mod https://forum.minetest.net/viewtopic.php?f=9&t=15423
03:12 paramat left #minetest
03:25 Rate85 joined #minetest
04:03 cheapie I wonder how many people are going to "get" these product names: https://cheapiesystems.com/media/images/screenshot_20160828_230255.png
04:04 cheapie (vending machines behind a car wash)
04:06 Miner_48er joined #minetest
04:21 est31 joined #minetest
04:27 Yst joined #minetest
04:51 est31 joined #minetest
05:20 namach joined #minetest
05:40 frostsnow joined #minetest
05:44 lumidify joined #minetest
06:00 red-001 joined #minetest
06:10 agaran cheapie: looks nice
06:11 * cheapie takes a picture of the whole building
06:12 cheapie agaran: https://cheapiesystems.com/media/images/screenshot_20160829_011124.png
06:12 cheapie The "vending machines" in the earlier picture are behind the automatic bay.
06:15 cheapie (oh, and yes, that automatic bay does work)
06:18 Nosrick joined #minetest
06:29 thepiercingarrow joined #minetest
06:29 thepiercingarrow hey, anyone on?
06:29 Nosrick I'm here, for now.
06:29 Nosrick What's up?
06:30 thepiercingarrow just curious, any reference for default?
06:30 thepiercingarrow im making a mod, want to know the exact names and properities of the stuff inside default, but cant find a list
06:30 Nosrick You've got two options; the Minetest dev wiki, or the lua_api.txt document in minetest/doc
06:31 thepiercingarrow they both explain the minetest object, and the modding API, but now the specs of the default modpack
06:31 Nosrick Oh!
06:31 Nosrick Just dig through the code.
06:31 Nosrick That's what I did.
06:32 thepiercingarrow ouch! okay then. xD didnt know i had to dig. thanks, then!
07:00 CWz joined #minetest
07:07 Darcidride__ joined #minetest
07:13 Dolus joined #minetest
07:21 IhrFussel agaran, Server thread spiked to 60+% a few times but whenever it happened the debug.txt doesn't show a thing...seems like I have to enable verbose and trace output
07:28 red-001 joined #minetest
07:34 AntumDeluge joined #minetest
07:40 IhrFussel With trace and verbose I feel like the log file will be larger than 1 MB in a matter of minutes haha
07:41 Trustable joined #minetest
07:57 misprint joined #minetest
07:59 Megaf Good morning
07:59 Megaf [07:12:39] <cheapie> agaran: https://cheapiesystems.com/media/images/screenshot_20160829_011124.png
07:59 Megaf cheapie: that is cool
07:59 cheapie Thanks!
08:01 Megaf cheapie: fluxbox :)
08:01 cheapie What about it?
08:01 Megaf you use it
08:01 Megaf I saw in your screenshots
08:02 Megaf cheapie: what software is generating this graphs? https://cheapiesystems.com/media/images/2015-08-07-213443_1280x1024_scrot.png
08:02 cheapie I used to.
08:02 cheapie I'm on awesome now.
08:03 cheapie rrdtool for the ones in the browser, gkrellm for the sidebar.
08:03 Megaf nice server case you have
08:03 Megaf xP
08:03 Megaf https://cheapiesystems.com/media/images/IMG_20151015_131415700_HDR.jpg
08:04 cheapie Yep, I have one server that's all nice and 3 that aren't.
08:05 Megaf imagine how many Raspberries you could fit in there https://cheapiesystems.com/media/images/IMG_20151102_145338767_HDR.jpg
08:05 Megaf and you have those giant capacitors, cool!
08:06 cheapie I have some giant HV-ish ones, and then a bunch of 350F@2.7V supercaps.
08:12 SylvieLorxu joined #minetest
08:14 Megaf anyone else getting <ERROR> 2016-08-29 05:10:51: [CurlFetch] servers.minetest.net/announce not found (Timeout was reached) (response code 0)?
08:21 agaran IhrFussel: can be, I saw 150lag today
08:26 IhrFussel agaran, since the restart?
08:28 agaran no , before
08:55 IhrFussel agaran, alright what's the current lag like?
08:58 agaran 1.5
09:05 shamoanjac joined #minetest
09:05 IhrFussel 1.5 is the step interval I specified so zero lag right now
09:06 agaran and it works well yes
09:06 Calinou 1.5 is a huge step interval
09:06 Calinou I suggest never setting it higher than 0.2
09:06 Markow joined #minetest
09:06 nosrick joined #minetest
09:06 nosrick Morning folks
09:07 IhrFussel Calinou, usually I have it set to 0.6 but I thought a higher server step interval would eliminate the lag...it didn't so it must be something about loading/saving contents
09:08 Calinou IhrFussel: does your server have a SSD? it can help
09:09 IhrFussel Calinou, the server provider states on the vserver page a SSD with 300 GB space yes
09:10 IhrFussel Also I already got to the point that it's only caused by the Server thread of the game not ConnectionSend, ConnectionReceive or Emerge-0
09:13 IhrFussel It is sudden spikes in CPU % usage...nothing constant, it happens at completely irregular times and the impact of the spikes is always different..sometimes the Server thread jumps from 5% to 60% for a few seconds, other times it only jumps to 30 or 40...but it's mostly below 10
09:18 IhrFussel And the most important fact: It did NOT lag back in May on the EXACT same hardware and exact same heavy mods ( I added a few light ones but the lags even occur with no mods enabled)...I'm strongly thinking it is something in the DB or map file
09:20 IhrFussel DB and map file are the same so DB*
09:20 agaran IhrFussel: 19.3 now
09:36 agaran IhrFussel: 17.
09:40 Szkodnix joined #minetest
09:43 Szkodnix joined #minetest
09:46 rubenwardy joined #minetest
09:51 Szkodnix joined #minetest
09:59 sonicpp joined #minetest
10:01 lumidify joined #minetest
10:06 IhrFussel agaran, --trace option only shows connection debug right? That is way too much...I think info and verbose streams are enough that's why I restart
10:09 xunto joined #minetest
10:17 namach joined #minetest
10:20 IhrFussel agaran, --info and --verbose seem to not be logged in debug.txt or I missed a conf setting...now I redirect the output to a verbose.txt file, should work nopw
10:21 segfault22 joined #minetest
10:21 segfault22 LinuxMint uses the SystemD spyware so I have to get rid of it
10:22 agaran IhrFussel: I never played with those so far... I thought that bookmarks are for teleporting but apparently not
10:23 lumidify joined #minetest
10:27 IhrFussel agaran, I never did either, but the output is so detailed now it should tell us what's wrong
10:28 segfault22 I opened the back-panel on my mini fridge and accidentally touched one of the transistor heatsinks while holding the panel, and it shocked me... is it bad when a transistor is leaking electricity out the heatsink?
10:29 PseudoNoob joined #minetest
10:40 xunto joined #minetest
10:43 eliotJ joined #minetest
10:54 lumidify Hi, is there a way for a mod/subgame to put all its custom world-specific save files into a separate directory inside the world directory?
10:54 Calinou I think that yes, you can put mod files in world dir
10:54 Calinou not sure how
10:55 lumidify I know how to put it into the world dir, but, to keep them separate from the rest, I want to put them into another folder inside the world folder.
10:55 lumidify Just for organization.
11:21 segfault22 you could make the folder a modpack, putting the mod folders in a single folder and in the single folder put a text file "modpack.txt"
11:22 segfault22 the text file can be empty, it just has to be there, then the mod-loader will treat all sub-folders as mod folders and iterate through them just like it would in the mods folder where the modpack folder is
11:26 est31 joined #minetest
11:29 FreeFull segfault22: What will you use instead? FreeBSD?
11:29 segfault22 Void Linux
11:30 FreeFull Can it run minetest?
11:30 segfault22 even though most of this software is only good until the end of electricity when earth returns to a state of balanced magnetics, I don't want SystemD
11:30 segfault22 well of course its a linux-based OS
11:31 FreeFull There definitely are linux distros incapable of running minetest
11:31 FreeFull At least not without heavy work
11:31 segfault22 it looks like its an entirely separate project, not a fork, so I am not sure
11:32 segfault22 they have a minetest 0.4.14 in their sources though, maybe they can support it
11:32 FreeFull Void Linux probably can run Minetest, assuming you can get irrlicht to work
11:32 segfault22 ahahaha irrlicht
11:33 segfault22 oh, well it looks like they have it too... no ha
11:33 Megaf segfault22: do they have a default DE?
11:33 segfault22 DE?
11:34 segfault22 means different things for different people, please specify...
11:34 segfault22 desktop environment ok
11:34 Megaf yep
11:34 FreeFull The wikipedia article says you can pick your DE
11:37 Megaf https://wiki.voidlinux.eu/Post_Installation
11:38 Megaf Void looks like what Arch used to be back in Arch Linux 0.7x#
11:40 red-001 joined #minetest
11:41 lumidify segfault22: I was actually talking about saving files in the world directory, so in worlds/myworld. You can do that, but I don't know how to organize it into different folders.
11:43 segfault22 I thought you were "talking" about putting mods in the world's (not worlds) folder and keeping them organized
11:43 segfault22 not really "talking" here, are we? is anyone?
11:43 segfault22 I have used that word wrong for way too long
11:43 segfault22 we are not talking, we are,... uh...
11:44 segfault22 I don't know
11:44 lumidify Fine, *typing* :P
11:44 segfault22 no typing is what you do to form the message, what we're doing involves sending a fully-typed message or messages
11:44 lumidify IRCing
11:45 segfault22 yeah
11:47 segfault22 but maybe its better to, instead of saying "talking" or "IRCing" when you are IRCing about someone IRCing something else, like "thought you were IRCing about <x>", better to say "thought you meant <x>"
11:48 lumidify Or you could say "who cares?" :P
11:48 segfault22 so you don't have to say was/were or about, and its shorter
11:48 segfault22 yeah, who cares
11:49 segfault22 "thought you were whocares-ing about <x>"
11:49 segfault22 < just kidding >
11:49 lumidify :P
11:52 segfault22 I was thinking about implementing an energy system mod for the game, but now I'm worried about losing the ability to produce electricity when earth reverts to a state of balanced magnetics so I have to create it in real-life too...
11:53 segfault22 you forgot to say "WTF"
12:01 Darcidride joined #minetest
12:23 shamoanjac joined #minetest
12:32 segfault22 I won't use Void Linux because there's too much crap to set up the desktop environment
12:34 rubenwardy lumidify:   minetest.mkdir(minetest.get_worldpath() .. "/folder_for_mod")
12:34 rubenwardy then eg:   io.open(minetest.get_worldpath() .. "/folder_for_mod/file.txt", "w")
12:35 lumidify Oh, thanks!
12:35 rubenwardy also, you should check that mkdir returns true
12:35 lumidify I was looking at ways to do it in Lua, but didn't realize that Minetest had a nice utility function.
12:35 rubenwardy if not minetest.mkdir(minetest.get_worldpath() .. "/folder_for_mod") then   error("Unable to create dir!")   end
12:35 lumidify *plain Lua
12:35 lumidify Okay, thanks
12:35 rubenwardy really the Minetest API should encourage this
12:37 lumidify Hmm, isn't minetest.mkdir documented on dev.minetest.net?
12:40 rubenwardy dev.minetest.net is not the offical documentation
12:40 rubenwardy it's community managed
12:40 rubenwardy https://github.com/minetest/minetest/blob/master/doc/lua_api.txt http://rubenwardy.com/minetest_modding_book/lua_api.html is the offical documentation
12:41 lumidify Okay, yeah, I guess I should have looked at lua_api.txt first.
12:42 Szkodnix joined #minetest
12:43 segfault22 not sure if unity-linux supports minetest and eclipse, or I'll have to make it all
12:43 segfault22 not sure if <x> or <y>
12:44 rubenwardy Linux using the Unity wm can run Minetest...
12:44 Calinou there's also the Unity game engine :p
12:45 segfault22 http://unitylinux.com/index.php?id=about
12:45 rubenwardy don't see why it wouldn't work
12:45 rubenwardy just need to get the dependencies
12:46 rubenwardy on older systems (like, 1998) getting libz or whatever to work may be a problem
12:46 rubenwardy I tried running Minetest an old laptop I had - Celeron processor, 16MB of RAM, 4GB hard drive
12:47 rubenwardy didn't work, unfortunately
12:47 agaran 16M ram? maybe in textmode it would work;)
12:47 Calinou libcaca back-end!
12:48 rubenwardy I could run this 2D city building game, I forget the name
12:48 rubenwardy not sim city
12:48 segfault22 I have a newer computer system that I built, inside an old beige ATX case - it has 8GB RAM GHz CPU/GPU hybrid (2-core) and 2TB hard drive, so no system limits are really a problem
12:48 rubenwardy you'd place roads etc, and police stations and fire stations
12:48 segfault22 3GHz
12:48 Calinou OpenTTD?
12:48 rubenwardy no
12:48 rubenwardy 1 sec
12:48 segfault22 my  key isnt working
12:48 rubenwardy I may still have the disc
12:48 segfault22 #
12:50 segfault22 yay
12:50 rubenwardy no can't find it
12:50 segfault22 aww
12:52 segfault22 maybe I should make my own Linux distro, that is entirely centered on Minetest
12:53 agaran Calinou: openttd is awesome
12:54 segfault22 it would have Minetest by default and updated via package manager, Eclipse LDT and Eclipse CDT for working with mods and the engine, GEdit with Lua syntax highlighting enabled by default, GIMP for making your textures and such, etc etc etc
12:55 LNJ2GO joined #minetest
12:55 lumidify GEdit? Why not Emacs?
12:55 Tux[Qyou] joined #minetest
12:56 rubenwardy lol
12:59 aix Emacs? Why not butterflies?
13:00 segfault22 emacs is hipster, cinnamon is hipster, <insert anything here> is hipster
13:00 lumidify Nah, I just set the universal constants, it's easier than the butterflies
13:01 segfault22 is Vector Linux "good"?
13:03 Darcidride_ joined #minetest
13:04 LNJ2GO joined #minetest
13:05 segfault22 oh my god, LINUX IS SKYNET
13:07 agaran vi, only vi..
13:07 lumidify Coming soon to Minetest: the editor wars!
13:08 agaran lumidify: my pick is more stone than yours;)
13:09 FreeFull kakoune
13:09 sonicpp joined #minetest
13:13 tpe joined #minetest
13:14 segfault22 Okay i'm going to use Puppy Linux
13:15 FurryFeggot joined #minetest
13:16 segfault22 yuck, furries
13:20 rubenwardy wtf is a furry anyway
13:26 betterthanyou710 joined #minetest
13:29 segfault22 rubenwardy, you dont wanna know
13:29 segfault22 but I have to tell you because we have to prevent them from taking over
13:29 rubenwardy it seems that it's just people that dress up as animals. I don't see wants wrong with that
13:30 nosrick They tend to sexualise animal traits.
13:30 segfault22 they get nasty, posting comic pr0n and vore everywhere, and they are violent against anyone who doesn't accept or support their beliefs
13:30 segfault22 and yes, what nosrick said, but they take it a lot further than that... I will not elaborate.
13:31 rubenwardy ah, well dressing up as animals is cool, but bestiality isn't
13:32 segfault22 well they can't have one without the other, its integrated into their belief system (kinda like sharia law and islam)
13:32 rubenwardy meh
13:33 rubenwardy I can see why a kid would want to be, say, a lion in a video game even if they're not a furry
13:34 arecascino joined #minetest
13:34 arecascino shamoanjac: nosrick: anything that needs my attention?
13:35 agaran rubenwardy: may I ask something 'bot lua/luajit?
13:35 rubenwardy not sure what you mean by "bot lua/luajit"
13:35 rubenwardy oh, about
13:35 rubenwardy sure, I guess
13:36 nosrick arecascino, there was an update last night. Bug fixes and content!
13:36 arecascino alright
13:37 agaran rubenwardy: I have something like filter-entries() that filters entries by criteria.. and I am unsure what is more efficient, pass criteria name and value as argument of function.. or pass function that evaluates to true/false so it is more flexible. Mostly I'd like to know if I lose much efficiency on passing function..
13:37 arecascino nosrick: what branch?
13:38 nosrick master
13:38 rubenwardy not using a function would be faster, but depending on context optimisation might not be needed
13:38 rubenwardy "premature optimisation is the route of all evils" or smth
13:38 nosrick Develop will update every few hours (probably), but is almost guaranteed to be a pile of shit
13:39 STHGOM joined #minetest
13:39 STHGOM joined #minetest
13:39 agaran rubenwardy: I know, but I have little clue about lua, in C, I would use objdump to see what resulting assembly is...
13:40 agaran it will be evaluating like 300-400 tables with 2+ entries each and that has to be quick.. (network rescan on edit/delete)
13:42 segfault22 I used to have Vista on the machine which I will be using for a server, except it was only available as 32bit and the machine is capable of 64bit
13:42 segfault22 so I switched to linux to get to use those other 32 bits
13:42 segfault22 4 bytes
13:43 agaran rubenwardy: but maybe you are right, do code, put measurement of execution time in there, check if you need to tweak
13:45 segfault22 is puppy linux good?
13:47 segfault22 is puppy linux good for minetest and eclipseLDT?
13:47 Preuk segfault22: does it fetch?
13:47 segfault22 xD
13:47 Preuk it's not worse than another one, it's just specifics there
13:48 Preuk if you're new to linux, i'd advise a more "mainstream" distro, i mean one with a big userbase and a lot of resources (Q&A, wiki, docs, etc.)
13:48 segfault22 I have used linux for some time but I have not modified it heavily...
13:49 segfault22 I am concerned about the SystemD spyware/backdoor situation (scope creep, my-way-or-the-highway attitude from devs when this was never the way before, opaque binary blob, etc)
13:49 Preuk thumb rule : if distro tag is on first page of tag list on http://superuser.com/tags it's good to go :)
13:50 Preuk well, for "pure" distros, it's a bit more confidential
13:50 segfault22 Since Ubuntu is using SystemD I have to also find a distribution that can be installed to run command-line only for a server machine
13:51 Preuk (hell, they removed most distro-specific tags)
13:51 Preuk ubuntu-server :p
13:51 Preuk arch is nice for that, as is gentoo (big user base, freedom of choice)
13:52 Preuk but you might sweat bullets for a few weeks before you feel home there
13:52 segfault22 lol
13:57 segfault22 I am seeing on the website some editions that are reduced to have no graphical interface, just the CLI
13:57 segfault22 but they are from 2013
13:59 agaran I have my debian without systemd.. but thats /me
13:59 lumidify joined #minetest
13:59 segfault22 cool
14:00 segfault22 however I am worried that as soon as some new package gets required for something else to work, said new package will depend on SystemD and force it to be installed without you being able to do anything else except just not update ever again...
14:01 sfan5 >I am concerned about the SystemD spyware/backdoor situation
14:01 sfan5 can you move your anti-systemd shilling back to /tech/?
14:01 segfault22 its not shilling
14:01 segfault22 I have good reasons
14:01 sfan5 spyware/backdoor is not one of them
14:01 rubenwardy Just don't let sof ar see them :)
14:02 rubenwardy without the space
14:02 segfault22 how can you prove that SystemD is not malware?
14:02 rubenwardy because it's open source?
14:02 sfan5 how can you prove that it is malware?
14:02 segfault22 we can't, we can only discern what it may be based on how it has behaved, and the behavior of the developers
14:03 sfan5 lmao
14:03 rubenwardy Same way you can prove Linux isn't malware
14:03 rubenwardy by linux I mean the kernel, but applies also to all areas
14:04 betterthanyou710 joined #minetest
14:04 segfault22 as long as Linus Torvalds is alive, we can rely on the kernel to be "safe"
14:05 segfault22 as in, not deliberately designed to be insecure
14:06 segfault22 upstart was fine and very useful, then one day the developers abandoned it and that's the excuse for switching to the new-and-updated SystemD
14:07 segfault22 upstart still works, but you can't find it in the package manager any more. For what other reason would they remove it if people may want to use it and its still usable, other than to prevent people from using it?
14:07 sfan5 they removed it because they didn't want to maintain it
14:07 sfan5 that simple
14:08 segfault22 the koolaid is good for some sheeple but I am not one.
14:08 sfan5 ok you can stop shilling now
14:09 segfault22 upstart was fine, there was no problem with maintaining it. SystemD is far more complex than upstart ever was, and the scope has creeped all over the place
14:09 behalebabo segfault22: apt show upstart
14:09 behalebabo apt show upstart-sysv
14:09 Preuk segfault22: do what real men do, switch to *BSD
14:09 segfault22 E: Unable to locate package upstart-sysv
14:10 behalebabo what OS version do you have?
14:10 * Preuk runs away laughting maniacaly
14:10 rubenwardy maybe sofar could pitch in here :D
14:10 segfault22 trying to get upstart-sysv from the archive, the package manager tells you that there are unsatisfied dependencies. trying to install those dependencies results in conflicts. Oh the version is linux mint 17.3
14:11 segfault22 I could just uninstall systemd without changing anything else, and hope it works
14:11 segfault22 else install puppy linux and get rid of it
14:12 behalebabo In Ubuntu 16.04 upstart-sysv simply requires a removal of the systemd-sysv and ubuntu-standard packages
14:13 behalebabo perhaps Mint is misconfigured
14:13 segfault22 perhaps
14:13 segfault22 I haven't configured anything related to the package manager, though, so it would be misconfigured by default
14:14 behalebabo Mint has it's own repositories as well
14:14 segfault22 yes, I have noticed that
14:15 segfault22 its too much problems, I'll just go ahead and use puppy linux
14:16 behalebabo There is also https://devuan.org/
14:16 Fixer_ joined #minetest
14:16 segfault22 cool
14:17 segfault22 if it has a minimal installation, for CLI only server, that will be great
14:17 behalebabo IIRC you can select to not install X during debian install
14:18 segfault22 cool
14:18 segfault22 devuan even has an .onion site for installing/updating through the tor darknet, c00l
14:18 behalebabo debian does as well
14:19 segfault22 would it be unreasonable to make a fork of minetest where it is configured to route all traffic through tor?
14:19 behalebabo yes
14:19 behalebabo that much traffic would be a great load on tor, and very slow
14:19 segfault22 it would be slow and burden the network, right?
14:20 segfault22 ok I wont do it
14:20 behalebabo stuff like torrenting or games don't work well with tor
14:21 segfault22 if SystemD is safe and all-good, why are so many people refusing to use it, why have so many developers left from projects that now use it, and why are there so many distributions that are designed specifically to prevent it being installed?
14:21 sfan5 segfault22: yes it would be unreasonable because it wont fucking work
14:21 sfan5 tor doesnt do udp
14:22 behalebabo there are difficult ways to do that, sfan5
14:22 behalebabo such as https://www.whonix.org/wiki/Tunnel_UDP_over_Tor
14:22 yusf[m] joined #minetest
14:23 Preuk 3000ms lag sounds great
14:23 sfan5 you need a vpn for that
14:23 segfault22 https://farm3.staticflickr.com/2195/2323907359_7ded67bccb.jpg
14:24 agaran Preuk: only ?;)
14:24 LNJ2GO joined #minetest
14:26 segfault22 TCP liquid Directions: 1. soak cotton wool with TCP 2. apply to affected area of patient 3. make sure it helped; if not, repeat. UDP liquid Directions: 1. open bottle 2. flick liquid in general direction of patient 3. bugger off
14:27 segfault22 is it not funny?... :C
14:27 Preuk :)
14:27 Preuk nice analogy
14:27 segfault22 :D
14:28 Preuk what about a PBEM net connector, like freeciv does ?
14:28 Preuk play 3 seconds, send mail, wiat for other players, retreive mails, play 3 seconds...
14:28 agaran Preuk: i'd rather use affected-blocks count.. than seconds
14:30 segfault22 is it reasonable to use a globalstep function to iterate through a list of "producer" machines connected to "consumer" machines, transferring some amount of power from each producer to the connected consumer(s) if there is any available (1 globalstep = every 0.05s)?
14:30 Preuk reminds me of a (pretty old) web game that synced by getting a static data file updated by server based on client events
14:31 Preuk segfault22: too fast, i'd advise you to use bigger "packets" with lower frequency
14:31 Preuk and use buffers on both sides
14:31 segfault22 buffers?
14:32 Preuk consumer gets a buffer that can run process for say 2-5 seconds
14:32 Preuk and consume only local (buffer) resources
14:32 segfault22 I was just going to use arithmetic to measure how much power should move from the producer's storage to the consumer's storage, is that a buffer?
14:32 Preuk transfer should then be done from producer to buffer with a lower frequency
14:33 Preuk depends on your need
14:33 Preuk for accurate simulation i don't know the engine weel enought to answer
14:33 segfault22 okay, should I try to make it run every other globalstep, or every fourth globalstep, or is there some other way to run the function at interfals greater than 0.05s (globalstep) and less than 1s (abm)
14:34 shivajiva joined #minetest
14:34 Preuk but for a "gamey" system, it's better to cheat and stremline process a bit
14:34 agaran segfault22: you need to find balance between physically accurate system and playability..
14:34 Preuk abm interval sounds good enough for any playable system
14:35 Preuk any high frequency system won't be noticeable (unless perfs drop) by players
14:35 segfault22 right, this won't be accurate to reality because magnetic resonance power systems don't exist yet
14:35 agaran if you dive into van der walls equations to simulate proper energy effort for steam compression to determine how much energy it actually stores.. you may get physics award recognition.. but not playable stuff
14:35 segfault22 xD
14:36 aix where do we schematics go?
14:36 segfault22 I'm just going to make a very simple power system, but I don't want to use an ABM (like technic) because it makes it look like the system is slow
14:37 sfan5 aix: you decide for yourself
14:37 Preuk segfault22: as i said, use high frequency simulation where it's noticeable (inside each component when monitoring)
14:37 Preuk and low granularity for general processing
14:38 Jordach joined #minetest
14:38 Preuk say you use ABM for power packet transfer between producers/processors/whatever
14:39 Preuk and a local buffer for each processor that can hold about 5sec worth of resource
14:39 STHGOM joined #minetest
14:40 Preuk updating this local buffer once per second (or 2 seconds ever) should be enough to keep it running smoothly, even if processing itself need 0.05sec steps for processing
14:41 Preuk best part is you can not even use globalsteps unless player is really close or even using processor form
14:41 segfault22 globalsteps are valid for loaded chunks, right?
14:42 Preuk there can bit quite a lot of loaded chunks
14:43 Preuk when setting up and tuning, it is indeed nice to have accurate mertrics, to fine tune and optimize everything
14:44 rubenwardy segfault22, globalsteps aren't per map block
14:44 Preuk but when your supply line is up and running, who care if internal steps simulation is updated every 10 seconds as long as final product keep flowing ?
14:44 rubenwardy global steps are run every dedicated_server_step (by default 0.1 s)
14:44 rubenwardy you may mean abm
14:45 segfault22 globalsteps are VALID for loaded chunks, right?
14:45 rubenwardy no, they have nothing to do with chunks
14:45 segfault22 ABM is still too slow
14:45 rubenwardy use the setting to make ABMs run more often - by default the min inteval is 1s
14:46 segfault22 what setting?
14:46 segfault22 interval can be only an integer representing seconds. anything less that 1 defaults to 1.
14:46 rubenwardy https://github.com/minetest/minetest/blob/master/minetest.conf.example#L1010
14:46 rubenwardy the interval is a float, it's just that the code by default only runs abms once a second
14:47 rubenwardy also, take a look at node timers
14:47 rubenwardy http://rubenwardy.com/minetest_modding_book/lua_api.html#nodetimerref
14:48 segfault22 I didn't know there was a node timer system, is that new?
14:48 segfault22 like it hasn't been around before 0.4.4
14:49 segfault22 I noticed there's a problem with using globalstep for the power transfer function - it will try to transfer power for nodes that are not loaded.
14:50 segfault22 node timer looks promising, but it regrettably won't be as simple as iterating through a list of producers connected to consumers.
14:51 rubenwardy a global step is equivalent to doing     function foo() minetest.after(0.1, foo) end; minetest.after(0.1, foo)
14:52 rubenwardy it runs a function once every server step
14:52 rubenwardy not per node or map block or map chunk or map sector or anything else
14:52 rubenwardy infact, minetest.after was implemented in lua using global step
14:52 segfault22 the timer will be implemented for producers, on_timer will iterate through the producer's list of connected consumers, first counting them and then transferring an amount of power based on the throughput of the conduit divided by the number of consumers...
14:53 segfault22 what you mean minetest.after() isn't called only once? That breaks my function to close my open files when the game starts, because its trying to close them forever
14:53 segfault22 new problem,...
14:54 segfault22 Is there a way to call a function only once after the game-server is initiated?
14:54 rubenwardy global step runs a function once every server step
14:54 segfault22 yes
14:54 rubenwardy so actually it's not equivalent due to the map shut down problem
14:55 segfault22 okay it doesnt have to be equivalent because I'm not going to use it, is there a way to run a function only once at the time the game-server is initiated (after mods load)
14:56 rubenwardy minetest.after(0, function()           end)    will run a function once, on the first ever server step
14:56 rubenwardy if you call it when the mod loads
14:58 segfault22 so minetest.after() doesn't run every globalstep or few globalsteps?
14:58 segfault22 if so then my code for closing the files is fine as it is, no change is needed.
14:59 segfault22 I will be using the node timer, thank you for telling me about it because I didn't know it exists
14:59 segfault22 not for closing my files, but for the power transfer thing
15:00 Yst joined #minetest
15:00 rubenwardy no, minetest.after runs a function once on the first server tick after the time elapses
15:01 segfault22 okay, thank you
15:04 lisac joined #minetest
15:05 KaadmY joined #minetest
15:06 Markow joined #minetest
15:11 Not_a_Robot joined #minetest
15:17 Void7 joined #minetest
15:20 whitephoenix joined #minetest
15:25 Freejack joined #minetest
15:30 M-JLuc joined #minetest
15:30 M-geir joined #minetest
15:30 sherkaton[m] joined #minetest
15:39 SylvieLorxu joined #minetest
15:42 Krock joined #minetest
15:42 Krock joined #minetest
15:47 IhrFussel agaran, it's a LOT to read...I only added the lines from just before the lags...only when CPU spiked to 60+% http://pastebin.com/9FkvH6c8
15:49 agaran IhrFussel: hmm.. looks like sending a lot of data to client plus bulk metadata updates?
15:49 red-001 joined #minetest
15:53 IhrFussel agaran, I'd more towards meta data changes...that would at least explain why it still happens after removing mods, after setting server steps higher, even after recompiling...meta data is saved in the map.sqlite file AFAIK
15:53 IhrFussel I'd lean*
15:54 Volkj joined #minetest
15:55 IhrFussel And yes I can confirm that the ENTIRE Server thread HANGS when it happens...cause there is not 1 single log entry at the time the lags occur
15:55 agaran yep, maybe its good to find how much metadata are there, or maybe trying to change database from sqlite to postgres if it uses indexing..
15:57 IhrFussel agaran, can you even migrate to another DB type with an existing world?
15:57 agaran I see no reason why not..
15:58 agaran if nothing else there is sql-translator.. it can port your db design from mysql to oracle/postgres and whatever else
16:10 IhrFussel Question to experts: Does the node meta data update on each server step? Or in realtime? I mean internally
16:13 red-001 isn't #minetest-dev the place for questions like that?
16:14 red-001 also what do you mean by update?
16:15 lumidify joined #minetest
16:16 IhrFussel red-001, like does the engine only read/write node data from/to the DB on each server step or also in between? For example when server step is set to 1.0 will it mean the node data in the DB is only updated once per second?
16:17 Krock joined #minetest
16:21 IhrFussel The actual problem: It seems like the lag spikes in my server are caused by many node_meta_data changes...so I'm wondering whether the engine actually stores all meta data changes in memory and updates the DB only after the next server step or if that happens anytime in between too
16:21 agaran IhrFussel: even if core stores data in database, it is not saved until sync is happening, though you can call fsync(), for that solution would be temporarly move db to some absolutely fast medium like shm, but SSD is not much slower..
16:22 agaran IhrFussel: other solution is use other database and see if problem persist..
16:24 IhrFussel agaran, no I'm wondering if my approach is just the exact WRONG direction...I thought increasing the server step interval would help reducing lag but maybe DEcreasing the interval would help IN CASE meta data changes only happen once per server step
16:24 agaran hmm that canbe true, and to find out one needs to find where metadata sync code is..
16:25 Ronsor joined #minetest
16:26 IhrFussel Cause if the engine only updates nodes + their meta data once per server step it would mean the game right now actually WAITS 1.5 seconds and gets maybe overloaded with the meta data changes before it can update everything and clear resources
16:27 alkotob joined #minetest
16:28 Krock PC shut down the 2nd time (within a week) for no obvious reason. Is this a bag sign? :(
16:28 agaran hmm all data are in blocks table.. metadata and everthing else..
16:28 Krock *bad
16:28 red-001 wouldn't the game seem laggier the longer time between server steps anyway?
16:28 agaran Krock: any odd noises?
16:28 Jordach Krock,check the power button on the front isn't getting stuck
16:28 Jordach or your temps
16:28 Krock agaran, none
16:29 agaran Krock: odd, for me such usuall was failing power supply or mainboard capacitors getting dead
16:29 Krock Jordach, already checked that one. A-ok. The temps are very low, measuring 42°C average temp for CPU
16:29 Jordach Krock, check power buttons connection to the mobo
16:29 red-001 check if the power cable is not loose properly
16:29 Krock hmm. Then it might me the latter one
16:29 Krock caps drying out
16:30 red-001 ignore the last word
16:30 Krock hehe
16:30 * Krock quit (Remote host closed connection)
16:30 IhrFussel To find that out I'll just reduce the server step interval to a minimum... red-001 , it updates the world actions slower yes but internally it should still accept any input...I'll just set it to 0.3 on next server restart
16:31 swift110 joined #minetest
16:31 Jordach Krock, no, not funny - quassel has arrows on part / join
16:35 alkotob_ joined #minetest
16:41 Ronsor joined #minetest
16:42 swift110 joined #minetest
16:49 Krock Jordach, and stars (*) for ACTION?
16:52 Telesight joined #minetest
16:54 Void7 joined #minetest
16:56 hmmmm joined #minetest
16:59 segfault22 is it okay to change the dedicated_server_step from 0.1 to like 0.05
16:59 segfault22 or even 0.01 or 0.005
17:04 segfault22 this wont be running on an old intel i960 or something, I'm going to host te server on a machine with a 2GHz AMD Athlon X2 64 processor, 160GB main hard drive and identical for backup, and 7GB RAM (even though the system was only designed to run a maximum 4GB RAM (possibly because it was the maximum available back then, and they thought "well then this is the physical limit to how much RAM you can put in it anyways, no sense saying it can support m
17:04 segfault22 ore, even though it really can" and technology marches on) it looks like it can handle the 7GB just fine
17:04 segfault22 THE
17:04 segfault22 damn KEYBPOARD
17:05 Calinou segfault22: sure, feel free to set it to 0.05
17:05 Calinou it'll make other player movement appear smoother for example
17:05 segfault22 okay good :D
17:06 Calinou there is not much point in going below
17:06 segfault22 tank you
17:06 segfault22 thank
17:06 segfault22 my H-key is messing up, so is the E key and both 3 keys...
17:07 LNJ2GO joined #minetest
17:08 Nosrick joined #minetest
17:11 segfault22 You know how when you turn on a laser-pointer (or an LED with a proper waveguide), the light goes out in a nearly perfectly straight beam, but when you put your finger over the opening it lights up the whole end of your finger? its kinda like how optical fibers work,... now if you take a whole bunch of those LEDs or lasers, and put them right next to a light-detector of the corresponding frequency, you make a sort of true optical keyboard, where th
17:11 segfault22 e light is only conducted to the light-detector when your finger is close enough to the LED/laser to reflect the light internally enough to reach where the detector is
17:13 segfault22 not nearly as complicated as making one of those light-based "keyboards" that projects an image of the "keys" onto some surface and detects when your finger is "pressing" one, but it would be more reliable (if tuned properly) so that it won't think your finger is pressing several keys just because its hovering over more than one
17:25 Krock got an "interesting" link in one of my new recieved possibly-spam mails: https://survey.andreldm.com/index.php?r=survey/index&amp;sid=1
17:25 Krock was anyone else here pointed to that site too?
17:26 rubenwardy not seen that
17:26 agaran neither me
17:26 Krock it seems that they mass-mail everybody who ever contributed to a popular open source project
17:26 segfault22 well I'm glad someone is interested in how C++ stuff gets developed
17:26 rubenwardy I get quite a few emails "offering jobs"
17:27 segfault22 but its not nice to mass-mail right? kinda like how its not nice to mass-destruct?
17:27 Krock yeah, me too. Another guy found my long-lost 6 million dollar and only wanted to know my credit card number
17:27 Krock seems legit
17:27 Krock *cough*
17:27 agaran lmao
17:33 segfault22 lool
17:34 segfault22 I'm getting this weird graphical problem in Eclipse, where a line of code is in the wrong place or some characters are doubled-up, except when I click on it or scroll away and back, it goes back to normal
17:35 octacian joined #minetest
17:35 rubenwardy ew eclipse ;)
17:37 octacian lol I'm currently having a discussion about inventories. It's getting quite heated... :D
17:39 Krock apt-get remove eclipse
17:39 Krock or was it uninstall?
17:39 Krock nvm, you see what I mean
17:40 Calinou joined #minetest
17:40 rubenwardy this will free 1.6 GB of space. Are you sure? [y/N]
17:40 agaran lol
17:41 rubenwardy I feel like there's a your mama joke in here somewhere
17:42 agaran not really, just amount of space that eclipse might indeed take is funny
17:43 segfault22 I might use something other than eclipse someday, that is if that something is ever having an interface as good as or better than eclipse, and depending on how easy it is to use compared to eclipse
17:44 rubenwardy I personally use Intellij, even though I have a feeling it's not completely foss
17:44 rubenwardy nice interface though
17:44 segfault22 currently, everything else has the advantage of taking less space, but setting it up is so much trouble, you're better off writing it onto the hard-drive platters with a magnetic pen
17:45 segfault22 oh and yeah, most of the alternatives aren't FSF-compliant
18:07 red-001 joined #minetest
18:07 namach joined #minetest
18:11 proller joined #minetest
18:32 Arcelmi joined #minetest
18:37 Void7 joined #minetest
18:50 twoelk joined #minetest
18:56 Calinou rubenwardy: IntelliJ/PyCharm Community Edition are not completely FOSS indeed - they have a few proprietary libraries
18:56 Calinou but the base code is Apache 2.0
18:56 Calinou obviously, the Ultimate Editions are proprietary
18:59 Nosrick joined #minetest
19:12 lisac joined #minetest
19:16 sythe joined #minetest
19:32 PseudoNoob joined #minetest
19:32 Hirato_ joined #minetest
19:35 Nosrick Got a quick question for anyone who has experience importing models into Minetest.
19:35 Nosrick Why is it rotated all weird?
19:39 exoplanet joined #minetest
19:49 Nosrick Never mind, turns out that .x is a crap format anyway
19:50 Calinou Nosrick: prefer .b3d, it is smaller
19:50 Nosrick Can't be edited, though
19:50 Nosrick I'm using .obj
19:50 KaadmY Nosrick: .x has major animation problems, too
19:50 KaadmY it did for me at least
19:52 Nosrick Thanks for the tip!
19:54 Calinou .obj has no animation support :P
19:55 KaadmY i think it can do armatures
19:55 KaadmY not sure how though
19:56 twoelk joined #minetest
20:07 segfault22 I have this function to look for a certain table within a table, based on the first value in the inner table, then find some other value in the inner table and return it. currently it uses ipairs and iterates through the whole outer table until it finds the inner table with the specified first entry. Is there a better way to do this?
20:08 swift110 joined #minetest
20:08 swift110 joined #minetest
20:11 Not_a_Robot joined #minetest
20:12 shamoanjac joined #minetest
20:15 sonicpp joined #minetest
20:16 Passant joined #minetest
20:20 * twoelk seems to have wrecked his firefox during updating
20:21 Passant hi there, i have someone who wanna know how to split stack using the android minetest client. since i only played on desktop 'til now i have no idea so i though i just ask for him here :)
20:36 segfault22 ipairs machine it is
20:40 Calinou joined #minetest
20:41 segfault22 is there a way I can make /dev/urandom fetch data from some atmospheric noise RNG site?
20:42 segfault22 better yet, if I get a quantum flux RNG device, connect it via USB, can I make /dev/urandom use it?
20:43 aix segfault22: hire some typists
20:43 segfault22 ...
20:44 Nosrick Anyone got any tutorials on world generation?
20:44 aix ...to write a PRNG for you
20:44 segfault22 nevermind, I'll go write my own OS with an opened hard drive and a magnetic pen now...
20:44 aix lol
20:44 aix and butterflies
20:44 aix don't forget that
20:44 segfault22 ... ... ...
20:44 aix though i guess you could use emacs
20:44 segfault22 xD
20:44 aix one of my favourite comics
20:44 sonicpp joined #minetest
20:45 segfault22 but all deh pr05 use the magnetic pen method
20:53 segfault22 the leather armies have prevailed, the Phantom Lord has never failed; smoke is lifting from the ground, the rising volume metal sound
20:53 Passant nevermind, i found an issue on github, thanks anyway
21:10 segfault22 I have found this problem where I can't specify a table key (the thing in square brackets) to take the value from a variable,... it treats the variable name as the key
21:11 segfault22 Lua is not the best after all
21:12 segfault22 the only way to force the key in square brackets [] to take a value from something else, is to make YET ANOTHER function which returns that value...
21:19 Trustable joined #minetest
21:22 behalebabo eh? can you give an example?
21:22 behalebabo you could also try wrapping the variable in parentheses
21:23 segfault22 okay
21:23 segfault22 maybe that will work
21:25 segfault22 there is something wrong with my code
21:26 segfault22 I was just using a different function for each possible key, as there were only 4 possible keys for the application
21:26 segfault22 maybe I should leave it as it is
21:30 segfault22 it works!!!
21:30 segfault22 Thank you!!
21:35 segfault22 behalebabo you helped me turn 36 lines of code into only 12! this is considerably more space-efficient, and elegant. Thank you :D
21:36 xunto joined #minetest
21:36 behalebabo np :)
21:38 segfault22 joined #minetest
21:40 Ronsor will someone play minetest with me, yet again? because i need to do something fun now cuz comcast boxes are obviously cheap
21:42 xunto joined #minetest
21:42 swift110 joined #minetest
21:45 segfault22 maybe you should host a server
21:50 DMackey joined #minetest
21:50 segfault22 there's also several servers with many people online, maybe you could have fun there
21:50 segfault22 oh and note that comcast is a scam/ripoff company, they have countless bad ratings on ratings sites from people they've screwed
21:52 Ronsor i do host a server
21:52 Ronsor and i know comcast is scam
21:52 Ronsor we also have att uverse
21:52 Ronsor which also sucks
21:52 LNJ2GO left #minetest
21:52 Ronsor gotta wait for that GFiber
21:52 Ronsor and my server is: ron-sor.net port 30000, default password: ronsorminetest
21:52 Ronsor so yes i host a server
21:53 segfault22 cool :D
21:54 Ronsor ever since 0.4.12 was relevant too
21:55 segfault22 you know the ISPs have been using terabytes/s fiber optic for a long time now, the only reason they are letting people have GBit speeds now is because they cheap out on network infrastructure, most notably the DSLAM/multiplexers, because they like to pocket most of the money people pay for the service instead of using the money to actually support and upgrade the infrastructure
21:58 Ronsor oh hey a forest fire
22:10 Ronsor how to stop a minetest forest fire
22:10 Ronsor ...
22:10 Ronsor http://prntscr.com/cbv68c
22:19 Ronsor well my town is dead
22:20 Calinou Ronsor: disable_fire = true
22:20 Calinou in minetest.conf
22:20 Calinou that won't undo damage though :P
22:28 Ronsor yeah
22:28 Ronsor tat's an issue
22:28 Ronsor its too late
22:28 Ronsor half the town is burnt
22:30 Ronsor also i found worledit
22:30 Ronsor can do that
22:32 paramat joined #minetest
22:34 DMackey- joined #minetest
22:45 Yst joined #minetest
22:46 Ronsor hi Yst
22:51 agrecascino joined #minetest
22:55 Yst Greetings, Ronsor!
23:02 Jordach !tell Krock yes
23:02 MinetestBot Jordach: yeah, sure, whatever
23:02 KaadmY :P
23:03 Jordach i'm reverse engineering old 1980's proprietary connectors due to them being unicorn rare
23:04 Jordach and having a suitable 3d printing replacement
23:04 Yst Hmm. It seems that MinetestBot has a bit of an attitude these days.
23:04 Yst Jordach: What do you need those connectors for, if I may ask?
23:04 Jordach Sony hifi equipment
23:05 Yst Hi-Fi is high-quality sound, right?
23:05 Jordach ye
23:06 Jordach ebay. their offical parts dealer doesn't even have it
23:06 Yst So how is the reverse engineering going so far?
23:07 Jordach http://freeservicemanuals.info/en/ <-
23:07 Jordach however, connectors aren't listed
23:07 Yst Hmm.
23:07 Jordach chips, part no. are useful
23:07 Jordach but the actual connectors (EBP-10) are 1.5v
23:12 Nosrick Anyone know what format the map generation heightmap comes in?
23:12 Nosrick Is it a 1D or 2D array? A table?
23:13 paramat flat array of x-rows joined end to end
23:13 Nosrick Aw, fuck
23:14 paramat so a table with a single index
23:14 Nosrick Is the index (x, z)?
23:15 paramat nope, (z - zmin) * chunksize + (x - xmin) + 1
23:15 paramat chunksize = 80 by default
23:17 Nosrick Ah, thank you.
23:17 paramat 'flat array' means a 2D space converted into a 1D table by joining the x rows end to end
23:18 paramat all 3D flat arrays (3D noise, voxelmanip) work the same way
23:23 lunacat7 left #minetest
23:26 Nosrick VoxelArea:new is returning nil.
23:26 Nosrick Why?
23:26 Nosrick local area = VoxelArea:new{ MinEdge = minp, MaxEdge = maxp }
23:26 Nosrick That's what I've got, it's nil.
23:28 FurryFeggot joined #minetest
23:31 paramat local area = VoxelArea:new{MinEdge=emin, MaxEdge=emax}
23:32 Nosrick My minp is emin
23:32 Nosrick And my maxp is emax
23:32 paramat are you using this inside an on-generated function?
23:32 paramat ah
23:32 Nosrick Yeah, I am.
23:33 paramat emin != chunk minp
23:33 paramat emin emax are the corners of the full voxelmanip volume which extends 16 nodes beyond mapchunk borders
23:34 paramat so vm volume is 112^3
23:34 paramat it's the mapchunk plus a mapblock-thick shell
23:34 Nosrick Right, okay.
23:34 Nosrick How do I get emin?
23:35 paramat essential because mapgen needs to generate stuff outside the mapchunk, and make stuff overlap
23:35 paramat local vm, emin, emax = minetest.get_mapgen_object("voxelmanip")
23:36 paramat do you need a simple example of lua mapgen?
23:36 Nosrick Maybe. Gimme a moment.
23:44 Tmanyo joined #minetest
23:45 paramat hm actually i need to write one, but my mods are good mapgen examples https://github.com/paramat?tab=repositories

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