Time  Nick          Message
08:14 mMegaf        Good morning ladies and gentlemen. How are we all today? Here it seems the good weather has ended.
08:23 mMegaf        😸
11:58 Fixer         "At the Build 2018 developer conference that's taking place these days in Seattle, USA, Microsoft announced support for custom JavaScript functions in Excel."
11:58 Fixer         ouch
12:08 Megaf         Fixer: because windows is already very secure
12:29 nerzhul       IhrFussel, steal value means the host OS take the CPU for other VM
12:29 nerzhul       IhrFussel, are you on AWS ?
12:30 IhrFussel     nerzhul, no, a German hoster but since steal is almost zero and load is still high it means it's my fault right?
12:31 nerzhul       load can be high if you have many open files
12:31 nerzhul       or if you are waiting for many I/O
12:31 nerzhul       load != cpu
12:31 IhrFussel     Let me check iotop
12:31 Mr_Pardison   \o
12:32 IhrFussel     IO% is max 5%
12:33 IhrFussel     It's a SSD, IO should not be the problem
13:03 Fixer         hey, little drum!
13:06 Mr-Pardison   hai Fixer
13:36 rubenwardy    Anyone know how to make the oom killer on Linux better?
13:36 rubenwardy    I've been writing high memory programs recently
13:37 rubenwardy    I'd like it to kill runaway processes before going into swap
13:37 rubenwardy    And maybe things like Firefox before that
13:37 rubenwardy    Maybe not
13:38 rubenwardy    Would be nice if it only put userland stuff into swap, and left the os responsive
13:38 sfan5         everything is userland
13:39 nerzhul       except kernel everything is in userland but you can change the oomkiler policy by playing in /sys :)
13:39 sfan5         if you want to prefer killing instead of swapping you need to disable swap
13:39 sfan5         setting vm.swappiness can also help
13:39 sfan5         also make your program set oom_score_adj on itself so linux will prefer killing it
13:40 nerzhul       vm.swapiness is just for offloading caches to swap
13:41 sfan5         oh it doesn't apply to programs?
13:42 nerzhul       no
13:42 nerzhul       it's just for kernel caches
13:42 nerzhul       if you want to kill swap reduce its size
13:42 nerzhul       generally you reduce the swapiness on VM because there is no interest to move disk caches to disk on a VM :)
13:42 nerzhul       especially on databases too
13:43 rubenwardy    s/userland/user/g
13:43 rubenwardy    I've written a kernel, I'm well aware of what userland is
13:44 nerzhul       rubenwardy, => https://lwn.net/Articles/317814/
13:45 rubenwardy    I could make a script to set all processes started on start up to -16 priority
13:45 nore          rubenwardy: ulimit -v is quite useful for high memory programs
13:45 rubenwardy    Although some of them probably shouldn't be that
13:45 rubenwardy    Oh, I'll look into that
13:46 rubenwardy    There's also gradle though
13:46 nore          not quite the OOM killer, but it will prevent the process from allocating too much (and will result in a memory error in that case)
13:46 rubenwardy    Since when did Android studio require more than 5GB, fucking hell
13:46 nore          just do "ulimit -v <memory in kB>" in your shell and it should work
13:46 rubenwardy    Also given that Google have brutally murdered material design just yesterday, Android isn't looking so great
14:48 ANAND         Is it possible to detect which face of a node is punched?
14:48 ANAND         using lua
14:50 rdococ        The time has come to yay
15:46 longerstaff13 'ello
15:47 ANAND         Hi!
15:53 Mr_Pardison   hai
15:56 Krock         \o/
15:57 ANAND         Krock, is it possible to detect which face of a node is punched using lua?
15:57 Krock         Yes.
15:58 ANAND         umm... how?
15:58 Krock         you have the pointed_thing table. use above and below to calculate the direction
15:58 Krock         and convert that direction to a facedir (API function available)
15:59 Krock         or don't convert it - depending on how you want to use it
16:00 ANAND         I want to place a node adjacent to the punched node
16:01 ANAND         which'll be placed on the punched face
16:01 ANAND         No pun intended :P
16:01 * Mr_Pardison ducks
16:02 ANAND         XD
16:02 Krock         even simpler then. place a node at pointed_thing.above
16:03 ANAND         what if the node was punched on it's side
16:03 ANAND         ?
16:03 Shara         above is relative to the direction you punched from
16:04 Shara         not literaly above along the y axis
16:04 ANAND         ah ok... thanks!
16:04 Shara         think of it like.. you point at the edge between two nodes.
16:05 Shara         so you then have .under as the node and .above as the side of that edge you interacted with it from
16:05 ANAND         .under is the node itself?
16:06 Shara         Check here: http://dev.minetest.net/pointed_thing
16:06 ANAND         .above and .under return nodes?
16:06 ANAND         ok
16:06 Shara         position of nodes
16:06 ANAND         oh, that's exactly what I want... ^
16:07 ANAND         Thanks!
16:07 ANAND         Also, is it possible to have an overlay texture for an air node?
16:07 Mr_Pardison   'ello Shara
16:08 Shara         Hi Mr P :)
19:23 redneonglow   hi
19:23 redneonglow   i wanted to know
19:23 redneonglow   for postgresql-based servers
19:24 redneonglow   what effect does the shared_buffers option have?
19:24 redneonglow   for a long time, until i noticed it last night
19:24 redneonglow   i had it set for the ridiculously low value of 18MB
19:24 redneonglow   i know it's minimum of 512MB with 50% of your memory reocmmended
19:24 redneonglow   which in my case was 18GB
19:25 redneonglow   so i made a typo when updating postgresql.conf from 9 to 10
19:25 redneonglow   9.x to 10.x
19:25 redneonglow   anyone here know what effect that would have had?
19:31 IhrFussel     redneonglow, I have no clue about postgresql but a simple google search revealed that 128 MB is the default size for that cache... 18 MB sounds very low then
19:33 redneonglow   IhrFussel:  the minetest thread about using postgresql said 512 MB minimum
19:33 redneonglow   i just want to know what effect it had
19:33 rubenwardy    more disc I/O
19:33 IhrFussel     I read 128 MB is the default size if you don't specify it...
19:34 IhrFussel     More I/O can slow down your machine
19:35 IhrFussel     redneonglow, install/run iostat and look at the I/O wait value
19:36 IhrFussel     On my machine (which runs for 100+ days already) the stats are like this:
19:36 IhrFussel     avg-cpu:  %user   %nice %system %iowait  %steal   %idle
19:36 IhrFussel     4,34    0,00    2,51    0,10    0,59   92,45
19:37 redneonglow   well, i was having some issues with that actually at some point
19:37 redneonglow   disk i/o
19:37 redneonglow   i dont know when
19:37 longerstaff13 is it hard disk or SSD?
19:37 redneonglow   it's on a vps
19:37 longerstaff13 ah
19:37 redneonglow   ssd i think
19:38 redneonglow   iostate reports values just below yours
19:38 IhrFussel     Mine is basically also a VPS but with dedicated resources
19:39 IhrFussel     if %iowait is around as low as mine then it's definitely not the issue
19:40 rubenwardy    they could also have no one in game
19:40 rubenwardy    so no blocks being loaded
19:40 IhrFussel     %steal tells you how much of your CPU time other customers stole from you
19:40 rubenwardy    the best test is to have 5 players fast flying around
19:41 IhrFussel     But iostat reports average since VM boot time
19:41 IhrFussel     redneonglow, what's your uptime?
19:43 IhrFussel     If your VM is only up a few hours/less than a day then the values will not be helpful
19:44 IhrFussel     Use "iostat 1" then to see realtime values every second
19:46 redneonglow   5 days 10:40
19:47 redneonglow   (the vps had to be rebooted last week for spectre/meltdown fixes)
19:48 redneonglow   avg-cpu:  %user   %nice %system %iowait  %steal   %idle
19:48 redneonglow   0.33    0.00    0.33    0.00    1.64   97.70
19:49 redneonglow   i cant tell you what it was like before i discovered the typo, i was having problems with abms but that was due to a buggy mod
19:49 redneonglow   and i fixed it
19:50 IhrFussel     iowait is a little high but that could be cause of the 18 MB cache you had
19:50 Mr_Pardison   c y'all tomorrow
19:53 redneonglow   IhrFussel: that i had the 4 days before this morning?
19:53 IhrFussel     iostat reports average values since boot time ... to be really sure you need to wait another day or 2 and see if %iowait goes down
19:53 redneonglow   oh ok
19:54 IhrFussel     My values for example rarely change cause of 100 days uptime and barely anything increases/reduces average load since boot
19:55 IhrFussel     decreases*
19:57 IhrFussel     Wait
19:57 IhrFussel     I think I misinterpreted the values cause of the formatting
19:58 IhrFussel     %iowait is 0.00
19:59 IhrFussel     %steal is 1.64 ... redneonglow check the values again
19:59 redneonglow   i was running iostat 1 before and just got the last one
19:59 redneonglow   2.24    0.00    0.29    0.04    0.79   96.64
20:00 redneonglow   %user   %nice %system %iowait  %steal   %idle
20:00 redneonglow   is the one i just ran
20:00 IhrFussel     Run just iostat again without the 1
20:00 redneonglow   that previous one was without 1
20:00 redneonglow   here's another
20:00 redneonglow   %user   %nice %system %iowait  %steal   %idle
20:00 redneonglow   2.24    0.00    0.29    0.04    0.79   96.64
20:00 redneonglow   the same
20:01 IhrFussel     Ok yes %iowait is near zero so all fine there
20:01 IhrFussel     The steal value is also nothing to worry about IMO
20:02 redneonglow   ok thank you everyone
20:02 IhrFussel     redneonglow, is your machine running slow right now?
20:02 redneonglow   no
20:02 redneonglow   i mean i can't really tell, it's a vps
20:02 redneonglow   the three minetest servers on it seem fine
20:03 IhrFussel     Check /status after joining the servers
20:03 redneonglow   no swap is in use
20:03 IhrFussel     If the lag values are low then all is OK
20:04 redneonglow   creative gardens (my slowest) is 0.66
20:04 redneonglow   digital farms is 0.86
20:05 rdococ        I want to create a YouTube channel, but only to specifically annoy them over how they keep shutting down educational channels
20:05 redneonglow   the valleys (my fastest) is 0.3
20:05 IhrFussel     I'd say a max_lag below 1.0 is fine on any server cause that means all internal delays never exceed 1 second
20:06 Krock         rdococ, make an educational video about why YouTube shuts them down
20:06 IhrFussel     redneonglow, is the slowest also the one with the most (heavy) mods?
20:07 rdococ        Krock: "Because they're assholes" then it gets shut down ;)
20:07 redneonglow   yes
20:08 redneonglow   actually under non-postgresql backends, and when i had 18MB, it sometimes went up to 5
20:08 IhrFussel     If you are interested in finding out which mods exactly cause the most lags enable the mod profiler
20:09 redneonglow   when i was having the problems with the abms (took care of that a few weeks ago with the mod profile) it was going ridiculously high to the point it was giving packet errors when you tried to connect
20:09 IhrFussel     ABMs should ideally run with a low interval + high chance...someone told me that once
20:10 IhrFussel     Cause that distributes the load between them more and less should fire at once
20:14 redneonglow   this was a case of an abandoned mod with bugs
20:14 redneonglow   i found an updated fork but it was causing crashes
20:15 redneonglow   so i looked at the code and fixed what i could
20:15 redneonglow   the bees mod (not to be confused with the bees in mobs_animal)
20:15 IhrFussel     Yes but there are very likely more ABMs running on your server and their interval is maybe high + low chance which causes more server load
20:15 redneonglow   you think that's the case now?
20:15 IhrFussel     MTG has at least 5 ABMs last I checked
20:16 redneonglow   i occasionally have abm warnings on creative gardens but that has LOTS of mods
20:16 redneonglow   but after i fixed the bees mod i almost never get the ABM warnings on farms
20:16 redneonglow   valleys doesnt have it
20:16 rdococ        Krock: Or maybe it's time to design an open-source equivalent (if one doesn't already exist) and switch to it
20:17 IhrFussel     Do you use mobs redo? The default interval is 30 I think there..I changed it to 15 + increased the chance value a bit
20:17 redneonglow   yes i use mobs_redo
20:17 redneonglow   i have to go later
20:19 Krock         rdococ, there's still IPFS and d.tube, whereas the first is opensource
20:27 Jordach       my eyes are now blind
20:33 rdococ        My nose runs and my feet smell... I was made upside down!
20:33 Calinou       can't hear you I'm sleeping!
20:33 Jordach       jesus christ what the fuck is this font rendering
20:34 Jordach       pixels so aliased i can cut myself with them
20:35 Krock         Calinou, one does not simply >>hear<< people over text based chat
20:36 Calinou       Jordach: my university's PCs have font antialiasing disabled for some reason (they're on Windows 7 which has ClearType enabled by default)
20:36 Jordach       >windows 7
20:36 Jordach       >not using MacOS
20:36 Jordach       just ask IBM about the cost savings
20:36 Calinou       there's some iMacs too which I use for the database class
20:36 Calinou       (Oracle DB in VMs…)
20:36 rdococ        >university not using Linux
20:37 Calinou       they do have Linux machines in some rooms
20:37 Calinou       I use them for the information security class
20:37 Krock         >defining usability by OS
20:37 Calinou       (Debian 8, so oldstable now)
20:37 rdococ        >university using oldstable
20:37 Calinou       olduniversity using stable
20:37 Krock         university oldusing stable
20:37 Jordach       >mfw you can run meltdown and spectre for free shit
20:38 rdococ        >olduniversity can't be upgraded in Linux RealLife 19.13 (the current stable release of real life)
20:38 Jordach       >gets 385mbit/sec
20:38 rdococ        > 38mbit/sec
20:38 rdococ        s/38/385/
20:38 Jordach       >only gets sub 80mbit/sec from a gameserver
20:38 Jordach       reeeeeeeeeeeeeeeeeeeeeee
20:38 rdococ        >keeps starting sentences with >
20:39 Jordach       >meme arrows
20:39 rdococ        <sworra emem
20:39 Calinou       I get 11/1 at home currently, with 28 ms best pings (pinging google.fr)
20:39 Krock         first world problems at its finest
20:39 Calinou       I enabled FastPath which helps with ping quite a bit
20:39 Calinou       (I got like 42 ms before)
20:39 Jordach       Reply from 8.8.8.8: bytes=32 time=22ms TTL=57
20:39 rdococ        Krock: do really rich people get zeroth world problems?
20:39 Calinou       64 bytes from 8.8.8.8: icmp_seq=3 ttl=57 time=25.6 ms
20:39 Jordach       :regional_indicator_o: :regional_indicator_o: :regional_indicator_f:
20:39 Calinou       I'm using 1.1.1.1 for DNS now too
20:40 Krock         rdococ, no because it requires an infinite amount of money
20:40 Jordach       reeee
20:41 rdococ        Krock: In a world where everyone has a supply of free money, would wealth be determined based on how fast their supply is?
20:41 Jordach       since i have a macbook, i have that damn stain of a U2 album in my iTunes libraryt
20:41 Jordach       rdococ: we hyperinflation now
20:41 rdococ        e.g. Alice gets £1000 a week, but Bob only gets £100, in a world where everyone gets money from thin air every week would those two people be considered of wildly different wealth?
20:41 Jordach       rdococ: economics says nay
20:42 rdococ        But unicorns say yeigh
20:42 Jordach       finite is great
20:42 Jordach       infinite gives the problem of inflation
20:42 rdococ        Jordach: but Alice will in general have 10x as much money as Bob
20:43 rdococ        after three weeks, even if you subtract £100 or so for utilities Alice would have £3000-£100=£2900 and Bob £300-£100=£200
20:43 Jordach       why do you think bitcoin is worth something yet shitcoins like ether are worth far less
20:43 rdococ        Alice would be able to spend more, or at least more rapidly
20:43 Jordach       one of these is finite
20:43 Jordach       and thus has a true value
20:44 rdococ        Jordach: my point was to wonder about a universe where what is valued is not money, but the rate at which you get it
20:45 Jordach       it wouldn't
20:45 Jordach       most people with a supply like that would never invest it and spend it frivolously
20:45 Jordach       it'd be measured on saving ability
20:46 Calinou       I'm stocking money these days since I don't have anything to spend it on
20:46 Calinou       :P
20:47 Jordach       invest it and watch your free money come back in time
20:47 Jordach       investing is better than savings accounts by far
20:49 rubenwardy    talking about Youtube shutting things down
20:49 rubenwardy    I'm still annoyed at them disabling adverts
20:50 rubenwardy    that's $5 less a year that I actually received because the adverts on my website exceed the threshold
20:50 rubenwardy    all from a kneejerk reaction
20:50 rubenwardy    People: Please ban Jake Paul
20:51 Jordach       welcome to youtube
20:51 rubenwardy    Youtube: *takes away coffee money from hundreds of people*
20:51 Jordach       if you have a tick that means youre their earners
20:51 Jordach       aka the whitelist of not having ads disabled
20:52 Jordach       if pewdiepie thinks about starting a competitor YT will die within 6 months
20:53 rubenwardy    I doubt that
20:53 Jordach       50m+ subs bring in a huge amount of ad money
20:55 Jordach       effectively, YT cannot afford him to leave
21:01 rubenwardy    woah
21:01 rubenwardy    just found out the company I'm starting at in a few weeks has a 24/7 gym and free breakfast + lunch
21:01 rubenwardy    and I was happy with just the salary
21:03 nerzhul       it's modern company, have nice things to keep employees at work more time than regular employee and pay them less
21:03 nerzhul       less taxes to have that than pay you
21:03 Calinou       "one of us!"
21:03 nerzhul       welcome to the modern world, you will see that in many companies
21:03 rubenwardy    it's a very competitive salary though
21:04 nerzhul       it doesn't mean you can't be paid more without less "privileges"
21:04 nerzhul       but yeah it's cool, especially for young employees :)
21:05 rubenwardy    they also have nerf guns apparently, not so keen on that
21:05 rubenwardy    seems a bit silly
21:05 Calinou       startups: "The future is foosball"
21:05 Calinou       :^)
21:05 rubenwardy    pRoFesSiOnAl wOrKiNg eNvIrOnMeNt
21:06 Calinou       the foosball table is a must-have. You need to do something while npm is installing packages!
21:06 rubenwardy    what is foosball?
21:06 rubenwardy    is it table football?
21:07 Krock         node modules vs data center. who's going to win
21:08 Calinou       yeah, table soccer
21:08 rubenwardy    ewww
21:08 Calinou       I met the world champion of tabler soccer a few years ago
21:08 rubenwardy    so football
21:08 Calinou       my mother interviewed him and everything
21:08 rubenwardy    soccer is only used by people who don't understand what football is
21:09 benrob0329    football is only played by people who don't understand what soccer is
21:09 rubenwardy    *american football
21:09 Krock         foot + ball = football.  hand + egg = handegg
21:09 Calinou       soccer is only played by people who don't understand what nexball is
21:09 rubenwardy    soccer is derived from social football or something
21:10 rubenwardy    but the US decided that they were special
21:11 Calinou       https://www.youtube.com/watch?v=aLSdq9UcSOQ
21:11 Calinou       *boing*
21:11 Calinou       I mean it.
21:11 rubenwardy    oh, nice graphics
21:11 rubenwardy    especially for a FOSS game
21:12 Calinou       Xonotic is still one of the best-looking FOSS games after all these years
21:12 rubenwardy    FOSS games always tend to have crappy graphics and evil developers
21:12 rubenwardy    <_<
21:12 rubenwardy    >_>
21:13 Calinou       :D
21:13 Calinou       Red Eclipse 2.0 will bring a nice visual upgrade to the game too
21:13 Calinou       (it'll use the Tesseract engine instead of the old Cube 2: Sauerbraten engine)
21:13 Calinou       it comes with 100% real-time lighting
21:14 rubenwardy    what's freeciv like?
21:14 Calinou       it's 2D, no?
21:14 rubenwardy    looks it
21:15 Calinou       you probably want 0 A.D. if you want a good-looking FOSS RTS
21:15 Calinou       (still in development)
21:15 rubenwardy    I found that to be very buggy
21:15 rubenwardy    and also didn't get what was happening
21:15 Calinou       "The Persian soldier said: SpiderMonkey exception at …"
21:15 rdococ        Xonotic looks like the kind of thing that would explode my computer
21:15 Calinou       I can run it at 400+ FPS :^)
21:15 * rubenwardy  installs
21:15 Calinou       there's a new feature which lets you see other people's framerates in the scoreboard
21:15 Calinou       I almost always have the highest FPS
21:16 rubenwardy    900MB??? :O
21:16 rubenwardy    that's like 20 Minetests
21:18 rubenwardy    arch needs better servers
21:18 rubenwardy    measly 8 MBps
21:18 Calinou       heh
21:18 rubenwardy    wait no
21:18 rubenwardy    it's already in mega
21:18 rubenwardy    so around 1 MBps
21:18 Krock         3600 Baud ought to be enough for everybody
21:19 rubenwardy    that's like 12x less than my apacity
21:20 Calinou       https://framapic.org/lBS5anfk4ps6/pMEhrABT92li.png
21:20 Calinou       I played a CTF game today :P
21:20 rubenwardy    that's not ctf! :O
21:21 Calinou       Xonotic CTF is pretty fun too :P
21:22 Fixer         Nexuiz
21:24 Fixer         so much shitposting
21:28 Fixer         rubenwardy: "dat slave cage has a nice gym"
21:39 rdococ        every while or so I hear a noise that's like "eh uh" but it sounds kinda like a bird,  but it's clearly an electrical device but idk where it's coming from
22:43 aerozoic      rdococ, https://www.thinkgeek.com/product/kmnr/