Time  Nick          Message
00:08 IhrFussel     When were formspecs added? Before 0.4.11 right?
00:35 Fixer         ^____^
00:36 Fixer         tenplus1 is on notabug
04:35 GreenDimond   minetest.decompress fails after re-joining the world
04:40 GreenDimond   when I compress data and decompress it while in the same 'session' it works fine
04:41 GreenDimond   but when I exit and rejoin, it fails
04:41 GreenDimond   "decompressZlib: inflate failed"
05:55 GreenDimond   no clue as to why it happens
06:11 rdococ        pffffghthhtt
14:45 MinetestBot   TenPlus1: Dec-12 04:27 UTC <paramat> please restrain offtopic on -hub. you've had many warnings and kicks but repeatedly and conciously break the rules, i'll move to temp bans soon
14:45 TenPlus1      hi folks
14:47 VanessaE      hey
14:48 TenPlus1      hi Vanessa, how's things?
14:48 VanessaE      eh, things just is. :)
14:48 TenPlus1      same :P
14:49 VanessaE      <insert off-topic stuff here just to irritate paramat>  ;)
14:49 TenPlus1      I've already been threatened with a ban, why not :P
14:49 TenPlus1      saw the mc furniture mod, didn't seem complete... just many colour of the same things
14:49 VanessaE      nothing's as complete as homedecor ;)
14:50 TenPlus1      heh very true :)
14:50 IhrFussel     I just upgraded my server cause it only had 250 MB space left... my hoster did CPU/RAM upgrade automatically...but I had to increase the partition size myself... i NEVER did something like that before and thankfully they offer optional GParted Live ... now I'm all set for another year or so
14:50 TenPlus1      popularmmo's did a video on it...
14:50 TenPlus1      hi fussel
14:51 IhrFussel     Hey there TenPlus1
14:52 IhrFussel     You know what's rude? They even charge me $20 JUST for upgrading and then they don't offer me to increase the hard disk to the new size ... I don't know, kinda shitty
14:52 TenPlus1      half assed support
14:53 IhrFussel     They say "you are responsible for your server, we don't help with that"
14:53 IhrFussel     "if you need help managing a server, choose a managed server instead"
14:54 IhrFussel     A managed server costs up to $50 more per month lol
14:54 TenPlus1      you'd think the least they would do was increase current partition size to new size... it's not tht hard
14:55 IhrFussel     It is extremely hard via command line... but thanfully I found that my server control panel had a virtual DVD-drive and they shipped it with a few pre-installed tools (that's how I found Gparted Live)
14:56 IhrFussel     With Gparted Live all I had a simple GUI and it was only 5 mouse clicks
14:57 TenPlus1      it's a handy iso to have around :D fixed many a pc using that one
14:58 IhrFussel     I wonder what they need those $20 for...probably just "cause they can"
14:58 TenPlus1      exactly, if they can they will...
14:59 IhrFussel     I doubt that increasing a KVM instance's resources is a difficult task to do
14:59 jas_          so how about that new debian?
14:59 jas_          hi tenplus1
14:59 TenPlus1      hi jas :)
14:59 jas_          \:D/
14:59 TenPlus1      new debian ? am already using latest kubuntu using deb
14:59 jas_          yes!
14:59 jas_          it's the best
15:00 jas_          (i have no idea, never used anything else.)
15:00 TenPlus1      ehehe
15:00 jas_          but i did install buster alpha 4, gonna see how that goes, on a headless pc.  mm, not using it yet tho
15:00 jas_          it's just sitting there, on.  whatchu been up to?
15:00 IhrFussel     I don't like the command to install packages... it's that "yum" crap or something...right?
15:01 TenPlus1      working over xmas and chilling :P
15:01 jas_          i dunno, i think it varies for different os's.  and some that aren't debian-based use apt, as well
15:01 jas_          oh yeah, gotta do it ten
15:01 jas_          r & r
15:01 TenPlus1      deb uses apt
15:01 jas_          it has super cow abilities
15:01 jas_          whatever that means
15:02 jas_          i didn't know for the longest time you could `apt' instead of `apt-get', that it was a frontend
15:02 jas_          i'm trying to do areastore with protector redo, tenplus1
15:02 jas_          it's a bit challenging for me, but fun anyway
15:02 TenPlus1      area store as in shops ? or saving positions ?
15:02 jas_          instead of checking all nodes in radius, their meta, for ownership
15:03 jas_          check against AreaStore
15:03 TenPlus1      oh right, just cacheing the owner/player names
15:03 jas_          i suppose so, you make it sound easy when you say it like that
15:03 IhrFussel     Oh wait...the issue I had with Debian is that it doesn't natively support PPAs
15:03 jas_          frankendebian?
15:03 TenPlus1      ppa's are an ubuntu thing mostly
15:04 jas_          that pr, that hybrid dog had, to speed up meta... i wonder how it would improve protector redo
15:04 jas_          it was never merged, or not thought to be a big improvement, i think.  or maybe it was?  i haven't checked over the holidays
15:05 TenPlus1      last pull for protector redo was areas hud compatibility, but doing that lagged servers
15:05 jas_          six million ways to lag a server, choose one
15:05 TenPlus1      hah yeah
15:06 TenPlus1      I tried adding a cache at one point and gave more issues, when to renew cache if protectors removed/changed hands etc. was easier to check live
15:06 IhrFussel     What's the default HUD update interval of 'areas'? I made mine 1.5 secs and it helps quite a bit
15:06 jas_          there's a lot of options, hard without a coordinated plan
15:06 TenPlus1      hi scr
15:06 jas_          brb
15:07 jas_          i'm back!
15:07 IhrFussel     It's also good that it doesn't update the HUD if the previous string of protections is the same as the current
15:07 jas_          it'll take me time and testing and i may fail
15:07 jas_          but it's fun to try anyway
15:07 jas_          i got it working in conjunction with meta check at the moment, but so many bugs and
15:08 TenPlus1      definitely, always worth trying... might find better ways of doing things :)
15:08 jas_          the way you say i almost should start over
15:08 jas_          i must have been raging earlier because i just started deleting newlines
15:08 jas_          mm, oh well.  it's hard to merge but i could try to put it back.
15:08 jas_          i should start over
15:09 jas_          it's great to have an in-game item to protect with
15:09 jas_          that it has to be crafted, and that there's a cost to ownership
15:09 jas_          the bigger the area, the higher the cost
15:09 jas_          so i always prefer protector redo over areas
15:09 Shara         Fussel, not sure why you moan about your hosting over that. It's pretty standard to just do that yourself.
15:10 IhrFussel     BTW I find it funny that all OS use the name "MBR/MS-DOS" for legacy partition tables
15:10 TenPlus1      lol, compatibility
15:11 IhrFussel     Shara, I can still find it rude to charge me $20 for that ... I could've destroyed everything on my disk and nobody would've helped me
15:11 Shara         Fussel, they are a business, and yes, unless you pay for managed services, why should they?
15:12 jas_          did you tell them that ihrfussel?  you should complain!
15:12 Shara         I had to do the same, and it went without problem, when I added more space
15:13 IhrFussel     I don't expect them to manage my server...but nobody told me beforehand that I would even have to adjust the partition size myself...I feel like they should've told me when they offered me to upgrade
15:13 Shara         Seems like common sense to me.
15:13 jas_          or experience
15:14 jas_          i wouldn't know that, i never ran a remote server before.
15:14 Shara         jas_: not really. I barely have any experience and I'm not even a Linux user outside of the MT server I run
15:14 jas_          hm, i guess so
15:14 jas_          i usually mess that stuff up personally
15:14 IhrFussel     Shara, I guess you also use a GUI tool to increase the size correct? Or did you learn to do it via command line?
15:15 jas_          use the ? key!
15:15 Shara         No, did it all via command line
15:15 jas_          i do remember that hehe
15:15 Shara         My provider was good enough to have guides you can check for help. No clue if yours has that
15:15 jas_          !title
15:15 MinetestBot   jas_: REACTING TO MINECRAFT 2 ANNOUCEMENT TRAILER!!! - YouTube
15:15 jas_          haha
15:15 TenPlus1      lolol
15:16 IhrFussel     Via command line was just too much for me... 1. Change boot order in control panel 2. Activate "Rescue System" 3. Login into virtual rescue instance 4. mount your hard disk 5. Check partition table type 6. Start fdisk and enter a bunch of commands 7. Write it to the partition table
15:17 jas_          now is remote linux distro installation/upgrade relating to an mt server on-topic?  it is, right?  i just want to make sure.
15:17 Shara         If it's too much for you, I get it. It's just that maybe you shouldn't moan about the provider when this seems pretty normal.
15:18 Shara         Ask around and see who can help if you need it.
15:18 IhrFussel     You also have to do all that "Rescue System" crap before you can actually change the partition?
15:18 * TenPlus1    inserts a Playstation 1mb memory card to upgrade fussel's server
15:18 jas_          just tell them to fix it for you!  you're a paying customer damn it!
15:19 Shara         It was a while back for me, so I don't remember all the steps. I know it wasn't 'fun', and it took more time than I would have liked, but it was perfectly possible.
15:19 TenPlus1      hi aerozoic
15:19 aerozoic      hey TenPlus1 !
15:19 IhrFussel     jas_, nah that's not how it works... I don't pay enough to get my server managed not even for 1 task... it's "pay less and do everything without us" or "pay a lot more and we do everything for you"
15:20 jas_          oh.  well then in that case, you get what you pay for.
15:20 jas_          !next
15:20 MinetestBot   Another satisfied customer. Next!
15:20 IhrFussel     I pay quite a bit... in fact after the upgrade it#s close to $50 monthly
15:20 jas_          that is a lot
15:20 jas_          but it's all relative.  there's books.
15:21 aerozoic      TenPlus1, quick question... Serveral players on my server have asked why there are no dogs. I was just wondering what ur thoughts were on that regarding mobs animals?
15:21 IhrFussel     The problem is ONLY really needed more SPACE... and they told me they cannot just increase my storage I have to upgrade to another tariff
15:21 jas_          IhrFussel: how is your server going anyway?
15:22 TenPlus1      we added wolves to xanadu using dmobs, can be tamed into dogs :D
15:22 jas_          did you see that 'whats wrong with ur server part 2' post about characters and classes?
15:22 IhrFussel     jas_, it's doing fine...now ... before the upgrade the map took almost all space and I had to constantly delete logs
15:22 aerozoic      oh cool thanks
15:22 jas_          i only made it one paragraph in so far
15:22 jas_          yeah logs man, just auto delete?
15:23 jas_          keep enough to see a day or two?
15:23 IhrFussel     Not enough... my space was 53 GB and the map is 48 GB ... so there was not much to spare
15:23 TenPlus1      sorry that should have been kpg mobs that adds wolves/dogs
15:23 jas_          also i think minetest.log is resource intensive when used a lot, is that true?
15:24 TenPlus1      but... if someone were to make some really nice doggie textures for the model I'll add to mobs animal :)
15:24 jas_          the map database? or that plus the logs?
15:24 TenPlus1      hi Jordach
15:24 jas_          yeah ten cannot blender i suppose
15:24 jas_          oh maybe he can
15:24 TenPlus1      blender confuses meh :P
15:24 aerozoic      TenPlus1, ok thanks!
15:25 jas_          maybe some day there will be a new pack of nice models for mobs
15:25 jas_          but i really love that dungeon master
15:25 TenPlus1      as they come out I try to replace... AspireMint has done a great job with spiders :)
15:26 jas_          that's awesome
15:27 TenPlus1      and we did the move from .x to .b3d models on some to save space
15:27 jas_          there's quite a few nice models spread throughout all the mods, really.  it'd be nice to see them all at once
15:27 jas_          (and of course, polycount, or whatever it is, that makes the client work harder to render scenes?)
15:29 Jordach       daily reminder that you can do cel shading in MT
15:29 jas_          jordach really made me jealous when he was talking about >60 fps minetest
15:30 TenPlus1      mobs_mc exists and has every single Minecraft mob added to the game :P
15:30 jas_          oh, yup
15:30 * Jordach     laughs in 250fps at 165hz display
15:30 jas_          yeah right 250
15:30 TenPlus1      hahaha, vsync limits me :P
15:30 jas_          singlenode
15:30 jas_          clouds disabled
15:31 jas_          ...do you really?
15:31 Jordach       https://i.imgur.com/6TYWVfB.jpg nice potato pc m8
15:31 TenPlus1      disable smooth lighting, clouds, fog, anticheat and rollback :D
15:31 jas_          yeah there's no way a stable 250 on a crowded scene with nodeboxes and shit and plus, it depends if you're moving
15:31 jas_          and try fov 110
15:32 Jordach       https://i.imgur.com/T107cuZ.jpg
15:32 Jordach       lmfao
15:33 Jordach       >tfw i7 8750H vs i5 4690k
15:33 TenPlus1      https://orig07.deviantart.net/057b/f/2015/104/f/5/survival_is_beautiful___minecraft_wallpaper__uhd__by_minecraftphotography-d818uiw.jpg
15:33 jas_          well i dunno, but i would like to test it.  because i use a low sensitivity and a lot of movement and a wide fov
15:33 jas_          so a constant, stable anything
15:33 jas_          would be pretty impressive
15:34 IhrFussel     jas_, just map.sqlite is 48 GB
15:34 jas_          that's huge
15:34 Jordach       jas_: i can double v_range and still have >60fps
15:34 IhrFussel     debug.txt gets already wiped on each restart (only keep last 3 million lines which usually means a few days back)
15:34 jas_          yes, i've gotten close on old hardware
15:35 jas_          so if you have new, i don't doubt it.  it's just, that constant butter
15:35 jas_          can you feel it?
15:35 Jordach       v_range is 230 in those screen shots
15:36 jas_          well minetest is not a fast paced action game
15:36 TenPlus1      I wish that mt map loading was faster though, so we could make proper dropper style maps
15:36 jas_          but sometimes it can be
15:36 IhrFussel     Now I have 8 cores and 24 GB DDR4 RAM ... what am I gonna do with that, I don't need it...I just wanted the extra space
15:37 jas_          so a good test is a video.  youtube supports 60 fps i believe
15:37 IhrFussel     4 cores and 12 GB RAM (before upgrade) were enough to run MT
15:37 jas_          and move around a love, the mouse
15:38 jas_          https://www.youtube.com/watch?v=BrjPDoY2kyM
15:39 TenPlus1      that npc is possessed
15:39 TenPlus1      also entity glitch in progress :D ahaha
15:40 Shara         Anyone know how I could find out who owned an area with a specific number, after the area was deleted? (I have backups of areas.dat, but it doesn't seem to record area number)
15:41 TenPlus1      prolly loads them in number order, counting each name as it appears
15:41 Shara         Yea, but the line I have in that position doesn't seem a likely candidate
15:41 rubenwardy    Shara: areas are stored in order
15:41 rubenwardy    Oh, inb4
15:41 TenPlus1      o/ ruben
15:41 rubenwardy    Hey TenPlus1
15:41 * rubenwardy  has been hacking areas recently
15:41 Shara         rubenwardy: but when you delete an area, then the next made takes it's number.... it's the ordering kept?
15:42 Shara         its*
15:42 rubenwardy    I think it populates gaps with null
15:42 rubenwardy    *nil
15:42 TenPlus1      then next entry replaces nil
15:42 Shara         Hmm, then maybe I need a more recent backup for this
15:42 IhrFussel     Shara, for that exact reason I log any deletion...to be able to track back the ID, owner and protection name ... your only hope is for i = 1, 9999 do print(areas.areas[i].." "..areas.areas[i].owner.." "..areas.areas[i].name) end << something like that
15:43 Shara         It's unlikely to be a major issue. But one of the team typoed when deleting an area and isn't sure what he hit
15:44 IhrFussel     Do you log commands? Then search for the ID that was typoed and see if it appeared somewhere else already
15:44 Shara         I don't think searching in general for a three digit number is going to be very helpful
15:45 IhrFussel     It depends... if someone added others to it then you could potentially see who's affected
15:45 jas_          they're handled internally unless you explicitly specify id, in insert_area()
15:45 jas_          so it could be any number, like sound handles
15:46 jas_          (i wish sound handles could be handled explicitly)
15:46 jas_          unless i'm mistaken.  i didn't actually try for a year+
15:46 IhrFussel     For example if someone protects something via /protect blah and later /add_owner ID someone_else blah ... then 'someone'else' could potentially know what it was
15:47 jas_          playing with it now tho, it's possible to get name, or "data" (i use player name) in get_areas_for_pos if you do `true' to return data
15:47 jas_          get_areas_in_area() returns data by default
15:47 jas_          odd
15:47 IhrFussel     All areas are stored in areas.areas ... you can loop through it like I did above
15:48 Shara         well, the issue is simply whether the latest backup I have of the file will be recent enough that nothing in the order was changed
15:49 IhrFussel     So stop server, replace areas.dat with backup one, start server and run a command that loops through every area and prints the info somewhere
15:49 Shara         Uhm, no
15:49 Shara         That would be a huge mistake
15:49 IhrFussel     I mean replace as in 'move them around'
15:49 TenPlus1      hi nerzhul
15:49 Shara         Fussel, big no
15:50 Shara         I'm quite capable of doing a search without needing to replace the actual up to date file
15:50 Shara         I'd lose a lot more than one area if I did that :P
15:50 jas_          ouch
15:50 IhrFussel     You could code your own lua table parser script and run it through that
15:51 Shara         As I said, quite capable of doing a search
15:52 IhrFussel     What kind of search? Searching the areas.dat won't work cause the IDs are not stored anywhere, the IDs are given by the order
15:52 Shara         ...
15:52 Shara         I'm not stupid Fussel
15:53 IhrFussel     So you need "some kind of script" that is able to parse the file ... there is no other way unless you try the search through logs method
15:53 Shara         Not at all
15:54 Shara         Split the file to get one entry per line and the line number gives me all I need, yes?
15:54 Shara         You just like overcomplicating things
15:54 IhrFussel     Not sure if that will work, you can try
15:54 Shara         It obviously works since it's a technique I use for all kinds of things
15:55 Shara         All I needed was to confirm how the file works. I don't need your help pulling data beyond that
15:57 TenPlus1      how big is the file ?
15:58 Shara         I completed the search about ten minutes ago
15:59 Shara         Even if Fussel thinks it's impossible :)
16:00 nerzhul       hi TenPlus1
16:01 nerzhul       #8000
16:01 ShadowBot     https://github.com/minetest/minetest/issues/8000 -- CSM restrictions: Make 'LOAD_CLIENT_MODS' disable loading of 'builtin' by paramat
16:01 rubenwardy    don't use IhrFussel's loop, it'll crash on any blank holes
16:02 Shara         There was no need for any loop
16:02 IhrFussel     That could be solved by checking "if areas.areas[i] then"
16:02 IhrFussel     I use such a loop to wipe abandoned protections and it works well
16:03 IhrFussel     I forgot to include that check in the snippet
16:04 rubenwardy    oorrr  do it properly
16:04 rubenwardy    for i, area in pairs(areas.areas)
16:04 jas_          that's clean
16:07 jas_          forgot the do tho
16:07 TenPlus1      lol
16:14 rubenwardy    I didn't forget it
16:15 rubenwardy    with that logic, I also forgot the end and the body of the loop
16:15 rubenwardy    :)
16:24 TenPlus1      hi tumeninodes
16:24 tumeninodes   Hey! Hi TenPlus1 with caps
16:24 TenPlus1      heh
16:28 tumeninodes   Hey anyone in the knowing... I was poking about things and came across Allow_ZWrite_On_Transparent in the irrlicht engine and wondering if MT utilizes this?
16:29 tumeninodes   I did not browse through enough to notice it
16:29 TenPlus1      am sure irrlicht has many kewl features we dont use :D
16:30 tumeninodes   oh, definitely
16:33 tumeninodes   I also came across a way to make the b3d loader cancel using mipmap, and I "think" it can be used for .obj too (uses a true/false value) which might give us less blurred meshes ingame?
16:33 tumeninodes   or use mipmap der
16:33 Amaz          Ooo, I just saw (and tried out!) the new pitch fly mode!! That's really cool (and fun!)
16:33 TenPlus1      hi amaz
16:34 TenPlus1      will try pitch fly soon enough :) gotta be smoother than the one we use in mobs api
16:34 rubenwardy    hey Amaz!
16:35 Amaz          Hi!
16:35 Amaz          Some very nice additions over the past couple of months :)
16:35 rubenwardy    it's so Amazing to see older players come back :)
16:35 Amaz          :P
16:35 * TenPlus1    awaits 5.0 stable
16:37 tumeninodes   does numberzero ever pop in here? would like to turn their attention to the 'allow_zwrite_on_transparent topic
16:37 Calinou       Quake had pitch fly 22 years before us :P
16:37 Calinou       I think even Heretic had it, IIRC
16:38 TenPlus1      wonder if we could add a swim/fly player animation in the mix
16:44 TenPlus1      hi benrob
16:47 benrob0329    hello TenPlus1
16:47 benrob0329    notabug was down yesterday, btw
16:47 TenPlus1      still is down
16:47 TenPlus1      oh wait, it's back up now... ahaha... was down earlier
16:48 benrob0329    nice
16:48 TenPlus1      wonder what the issue was
16:48 * benrob0329  is finally playing MT again
16:49 benrob0329    light survival with friends, mostly just building
16:49 TenPlus1      :P
16:50 TenPlus1      added vanessa's autocrafter to xanadu and made it work with hoppers :) players going nuts building crafting machines
16:50 TenPlus1      am considering adding sorting tubes
16:51 Calinou       TenPlus1: surely, but someone would have to create one… and it'd be very difficult to make a convincing flying animation since we can't change it based on the direction easily
16:51 Calinou       (in a way that looks good)
16:51 Calinou       even in Minecraft, they kind of gave up with a proper flying animation IIRC
16:52 TenPlus1      we have a laying down, just fliup it to looking down :D
16:52 benrob0329    Calinou: they have elitra with a proper flying animation now
16:52 benrob0329    looks pretty good tbh
16:53 TenPlus1      am sure pitch and yaw rotation would help make it look nice :)
16:57 Calinou       benrob0329: I mean, in creative mode
16:57 Calinou       TenPlus1: if it's interpolated, sure
16:58 TenPlus1      wb src
17:00 tumeninodes   https://github.com/svn2github/irrlicht/blob/master/branches/shader-pipeline/include/SceneParameters.h
17:00 tumeninodes   !title
17:00 MinetestBot   tumeninodes: irrlicht/SceneParameters.h at master · svn2github/irrlicht · GitHub
17:01 TenPlus1      laters folks o/
17:01 tumeninodes   he leaves so fast
17:47 GreenDimond   Does anyone have any ideas as to why minetest.decompress is failing?
17:56 Krock         o/
17:56 GreenDimond   \o
17:57 GreenDimond   Krock: Whatever happened to fixing the entity attachment collisionbox bug? :P
17:59 Krock         GreenDimond: fixed
17:59 Krock         at least for first-level attachments
17:59 GreenDimond   I mean the bug where they dont work outside of 200m around spawn
18:00 IhrFussel     What is 'legacy_mineral' inside a node definition? Needed by some mod?
18:00 Krock         nothing. dunno how to fix and spent not much time on it
18:00 Calinou       I think it's from the 0.3 days
18:00 GreenDimond   Didn't paramat give you the answer a while back?
18:00 GreenDimond   The camera updates every 200 nodes
18:01 Krock         GreenDimond: was it really paramat? I know about the camera behaviour but I didn't investigate further
18:01 Krock         interestingly it's the collisionbox AND selectionbox which isn't working properly
18:01 GreenDimond   I think he was the one that mentioned the camera behavior
18:01 GreenDimond   well yeah
18:01 IhrFussel     Calinou, was that directed @ me?
18:01 Krock         so I'm not sure how much it's related to the camera
18:02 GreenDimond   IhrFussel: Probably
18:02 Krock         IhrFussel: yes
18:02 Krock         just remove these old parameters
18:02 GreenDimond   I really think it would be worth fixing before 5.0 IMO
18:02 IhrFussel     So... drop = 'samemod:sameitem' is also legacy? You had to define the drop for nodes in the past?
18:02 GreenDimond   No
18:02 GreenDimond   well
18:02 GreenDimond   maybe
18:03 GreenDimond   drop isnt legacy
18:03 IhrFussel     No but I wonder why the node def of that item defines "drop" as itself
18:03 GreenDimond   If it's an ore then maybe they didn't want it to drop a lump
18:03 IhrFussel     Nope it's the 'falling anvil'
18:04 GreenDimond   except that wouldnt make sense anyway because ores drop themself unless drop defines something else
18:04 rdococ        thoughts: hmm
18:04 GreenDimond   Krock: Well unless something else updates every 200 nodes, camera makes sense
18:05 GreenDimond   Also minetest.decompress is being a pain
18:05 GreenDimond   A pointer on that would be great if anyone has one
18:06 IhrFussel     Uh weird...I can't even find that mod anymore it must be super old
18:07 IhrFussel     I found it!
18:07 IhrFussel     https://github.com/Megaf/anvil/blob/master/init.lua
18:07 * Krock       give GreenDimond Decompressor *dec_ptr = nullptr;
18:08 GreenDimond   Gee, thanks
18:08 BillyS        Can entities (other then players) have inventories?
18:09 IhrFussel     Now that I think about it the item is by Megaf and AFAIK they don't know much about MT modding
18:09 BillyS        It doesn't seem that they can ... that's quite annoying
18:10 Calinou       IhrFussel: yeah
18:10 IhrFussel     In case I remember correctly and Megaf was the server owner who couldn't really fix his mods and blame core devs for it
18:11 GreenDimond   BillyS: Use detached inventories
18:11 BillyS        I guess not; that seems to be an ongoing issue here: https://github.com/minetest/minetest/issues/6081
18:11 GreenDimond   There are plenty of NPC Trader mods
18:11 GreenDimond   take a look at them
18:11 BillyS        GreenDimond: That's gonna be my only option, but saving the inventories over reboot is gonna be something horrible
18:11 GreenDimond   Nah
18:11 GreenDimond   Easy
18:11 GreenDimond   Staticdata aint horrible
18:12 GreenDimond   minetest.serialize(inv:to_table())
18:12 BillyS        ...derp
18:12 BillyS        Didn't think of that xD
18:12 GreenDimond   As long as your inventory isnt 1000 slots
18:12 GreenDimond   If it is then you could minetest.compress it
18:12 GreenDimond   If you can get decompress to work, that is >:(
18:12 BillyS        Nah, it's more like 12
18:12 BillyS        Okay, thanks
18:16 Krock         BillyS: entities cannot have inventories. serialize it manually
18:17 BillyS        So it seems, thanks
18:17 Krock         (just to confirm what's been said above)
18:17 Krock         (because I didn't read it)
18:18 GreenDimond   Krock: any chance of the bug being worked on before 5.0?
18:18 Krock         yes, I'll make the bug better
18:18 Krock         xP
18:18 GreenDimond   Thanks x2
18:19 GreenDimond   Your null pointers and bugs are so helpful
18:20 Krock         I'll give it a try - but the camera code is overcomplicated
18:20 GreenDimond   Thanks :)
18:20 GreenDimond   Now I need to figure out why minetest.decompress is failing
18:20 GreenDimond   and I dearly hope it isnt an engine bug
18:20 GreenDimond   which it probably is
18:21 Krock         failing with 5.0.0-dev?
18:21 GreenDimond   Yes
18:21 Krock         error dump pls
18:22 GreenDimond   literally "luaentity_Activate(): decompressZlib: inflate failed"
18:22 GreenDimond   (it is being called when an entity is activated)
18:22 GreenDimond   and it shouldnt be a problem
18:23 Krock         if(status == Z_NEED_DICT || status == Z_DATA_ERROR
18:23 Krock         || status == Z_MEM_ERROR)
18:23 Krock         corrupted data
18:23 GreenDimond   When the entity activates it runs fine
18:23 GreenDimond   but when I reload the world
18:23 GreenDimond   the entity activates and decompress fails
18:23 Krock         add some sanity checks to ensure it's not an empty string and the correct length
18:24 GreenDimond   correct length?
18:24 Krock         i.e. minetest.serialize({ length = string.len(comp_data), data = comp_data }
18:24 Krock         on read the "data" value but be the same as "length"
18:25 GreenDimond   that doesnt make sense though
18:25 GreenDimond   you serialize before compressing
18:25 GreenDimond   minetest.compress takes a string
18:26 GreenDimond   are you saying serialize again after compressing?.
18:30 Krock         GreenDimond: attachments work when commenting ou the camera shift
18:30 Krock         GreenDimond: yes, to pack it into a proper string/table
18:31 GreenDimond   So it is the camera causing the issue?
18:31 Krock         need to verify
18:32 Krock         cannot reproduce in the vanilla client either
18:32 Krock         same testing code. what is happening huh
18:33 GreenDimond   Cant reproduce what?
18:33 Krock         the 200m attachment bug
18:34 GreenDimond   erm
18:34 Krock         lol
18:34 Krock         broken in an older build
18:34 GreenDimond   But
18:34 Krock         maybe it's accidentally fixed by my recent camera changes
18:34 GreenDimond   my build is from like a day ago
18:34 Krock         I just noticed my "vanilla" build is not that vanilla
18:35 GreenDimond   :v
18:36 GreenDimond   alright now im confused
18:36 GreenDimond   attempt to index local "dat" (a nil value)
18:36 GreenDimond   on the line right after its defined
18:36 GreenDimond   wtf
18:36 GreenDimond   local dat = minetest.deserialize(storage:get_string(self._id))
18:37 GreenDimond   its failing on the next line, if dat.data == "" then error("Data is empty") end
18:39 BillyS        GreenDimond: Actaully, it seems that inventories don't have a to_table method, although MetaDataRefs do
18:39 Krock         GreenDimond: fixed by #8015
18:39 ShadowBot     https://github.com/minetest/minetest/issues/8015 -- Camera: Code improvements, use Raycast by SmallJoker
18:39 GreenDimond   where is MinetestBot's api command when I need it
18:39 Krock         (attachment bug)
18:39 Krock         !commands
18:39 BillyS        https://dev.minetest.net/InvRef
18:39 Krock         !help
18:39 MinetestBot   https://github.com/sfan5/minetestbot-modules/blob/master/COMMANDS.md
18:40 Krock         GreenDimond: ID is always different dude
18:40 Krock         assuming you're using the internal entity ID
18:40 GreenDimond   I am not
18:40 Krock         and it must be a string
18:40 GreenDimond   Ignore the storage part
18:40 GreenDimond   It works
18:40 GreenDimond   100%
18:41 GreenDimond   Because I used this with just minetest.[de]serialize
18:41 Krock         is it empty after the .compress call?
18:41 GreenDimond   The string is not empty, I at least confirmed that
18:41 GreenDimond   Remember Krock, it works in the same session
18:42 GreenDimond   When the entity is created and it activates, compress and decompress work fine
18:42 GreenDimond   it's when I exit and rejoin that it fails
18:42 GreenDimond   wtf
18:43 GreenDimond   no matter what I do it cant find the local var
18:43 GreenDimond   is that because deserialize is returning nil?..
18:43 GreenDimond   I bet thats why
18:43 Krock         you must define locals above where you use them
18:43 GreenDimond   ...
18:43 GreenDimond   Im not stupid
18:43 Krock         no really you might expect "foo" to exist inside a function
18:43 Krock         since the function is by time called after the main code body
18:44 GreenDimond   It is inside the proper scope
18:44 Krock         do you have the entire relevant code somewhere to test it?
18:44 GreenDimond   No
18:45 Krock         oh wait
18:45 Krock         got an idea
18:45 GreenDimond   somehow its getting nil
18:45 Krock         what if you use set_string and get_string after each other?
18:45 Krock         is get_string returning an empty string?
18:45 Krock         binary data may contain a leading zero
18:45 GreenDimond   Looking at the mod storage file it gets writen
18:45 Krock         causing the string reading function to fail
18:45 GreenDimond   written*
18:45 Krock         yes szre
18:45 Krock         it's about reading
18:45 GreenDimond   ah
18:46 GreenDimond   let me undo the sanity checks
18:46 Krock         seen that issue before, but didn't remember in which context
18:46 Krock         std::string name = luaL_checkstring(L, 2);
18:46 Krock         boom
18:47 GreenDimond   how do I check the get_string
18:47 GreenDimond   dump or print I suppose?
18:47 GreenDimond   I was about to do chat_send_all but it isnt a standard string so I figured that wouldnt go well
18:47 Krock         storage:set_string("test", minetest.compress("your data here"))
18:47 GreenDimond   ?
18:47 Krock         storage:get_string("test") then compare
18:47 GreenDimond   wait
18:48 GreenDimond   compare how?
18:48 GreenDimond   nvm
18:48 Krock         err == and length
18:49 Krock         or write it to a file to ensure it's not stopping at \0
18:49 GreenDimond   if I do local compressed = bla; storage:set_string(id, compressed); local get = storage:get_string(id); minetest.chat_send_all(tostring(get == compressed))
18:49 Krock         yes
18:49 GreenDimond   alright
18:50 Krock         but compressed must be your binary data
18:50 GreenDimond   binary being after minetest.compressed?
18:51 GreenDimond   returns true
18:51 GreenDimond   on first gen
18:51 GreenDimond   let me see on reload
18:52 GreenDimond   wait
18:52 GreenDimond   nvm
18:52 GreenDimond   lol
18:52 GreenDimond   It doesnt compress anything on activate
18:52 GreenDimond   Krock: So yes it is writing it properly
18:54 Krock         GreenDimond: are you "GreenXenith" on Github btw?
18:55 GreenDimond   Yes
18:55 GreenDimond   the mod isnt there though
18:55 Krock         writing a test file to reproduce this
18:55 GreenDimond   ah
18:57 Krock         GreenDimond: is the ID alphanumeric only or could it contain the null character?
18:57 Krock         doesn't really matter though
18:57 GreenDimond   its a number
18:57 Krock         ah. so it's converted to a string by Lua
18:57 GreenDimond   um
18:57 GreenDimond   I guess?
18:57 Krock         yes. it was a statement. sorry :3
18:58 GreenDimond   That wouldnt be the problem would it?...
18:58 Krock         no, just trying to eliminate all possible causes
18:58 GreenDimond   I can test the attachment thing with my own mod btw
18:58 GreenDimond   let me just pull the latest build
18:59 Krock         afaik there were no changes in this code area
18:59 GreenDimond   already up to date anyway :x
19:00 GreenDimond   collisionboxes still dont work outside of 200
19:00 GreenDimond   *boxes *
19:04 IhrFussel     Is num_emerge_threads = 7 still ok or too much?
19:04 GreenDimond   Krock: Want the binary line?
19:04 IhrFussel     I have it at 3 but now I got upgraded from 4 -> 8 cores
19:06 Krock         GreenDimond: try my PR in the meanwhile .. still working on the test script. slow fingers
19:06 Krock         IhrFussel: minetest only uses one core anyway
19:06 GreenDimond   yuck I have to figure out how to test a PR again
19:06 GreenDimond   lovely
19:06 IhrFussel     Krock, no, the setting is for map generation and it allows to use more than 1 I thought
19:06 Krock         curl -Ls <pr url>.patch | git am -3 -
19:07 Krock         IhrFussel: it allows to set threads, not cores
19:07 IhrFussel     But I read everywhere that more threads means more cores are being used
19:07 Krock         *shrug*
19:08 IhrFussel     "On multiprocessor systems, this will improve mapgen speed greatly"
19:09 IhrFussel     Not sure if multiprocessor only means several separate CPUs
19:10 Krock         GreenDimond: funny how "method" in minetest.compress is not even used
19:10 GreenDimond   what?
19:10 GreenDimond   yes it is
19:12 IhrFussel     Okay so..num_emerge_threads is actually pretty useless to increase on a multiCORE CPU?
19:13 Krock         GreenDimond: serialization issue is due to either writing to the file or reading from it
19:13 GreenDimond   compression*?
19:14 Krock         not compression
19:14 GreenDimond   but decompress is the thing failing
19:14 GreenDimond   not deserialize
19:18 GreenDimond   Krock: Bug not fixed with PR applied
19:18 GreenDimond   wait
19:18 GreenDimond   derp
19:18 GreenDimond   need to rebuild
19:18 GreenDimond   iamverysmart
19:19 GreenDimond   :(
19:19 GreenDimond   Krock: No dice
19:20 tumeninodes   https://www.youtube.com/watch?v=NVvMQ4WAOAo&feature=youtu.be
19:20 tumeninodes   plucking at glass use underwater
19:21 Krock         "no dice" = "no success"?
19:21 GreenDimond   Yes
19:21 GreenDimond   *boxes still not there
19:23 Krock         regarding the serialization issue, somewhere 0x40 is added to the value
19:23 Krock         I have no idea why
19:23 GreenDimond   So is this an engine bug?
19:23 Krock         yes
19:24 Krock         it's basically byte |= 0x40; somewehre
19:24 GreenDimond   oof
19:24 GreenDimond   btw
19:24 GreenDimond   even your test script doesnt work
19:24 Krock         uh? I tested it multiple times locally
19:25 GreenDimond   did you test outside of 200?
19:26 GreenDimond   wait
19:26 GreenDimond   my window version doesnt have -dirty
19:26 GreenDimond   but it built -dirty
19:26 Krock         wrong binary file
19:26 GreenDimond   ..
19:26 GreenDimond   but
19:26 GreenDimond   no
19:26 GreenDimond   how
19:27 GreenDimond   it's the correct file
19:27 Krock         https://i.imgur.com/Z9c70g6.png
19:27 Krock         oh it's only dirty when git reports a non-clean workspace
19:28 GreenDimond   well it detected dirty when building
19:29 GreenDimond   hm
19:29 GreenDimond   Yeah its still not working
19:30 Krock         Works On My Machine (TM)
19:30 GreenDimond   I dont get why the window doesnt say -dirty
19:30 GreenDimond   when building it says Detected Git version 5.0.0-dev-f12d3749-dirty
19:31 GreenDimond   erm
19:31 GreenDimond   it isnt binding the executable
19:31 Krock         if the git hash matches with HEAD then it's fine
19:33 GreenDimond   Time to reclone
19:33 Krock         not clone. apply the patch and make
19:33 GreenDimond   I already did Krock
19:33 GreenDimond   and it didnt do it right
19:33 Krock         not making properly :3
19:34 GreenDimond   so this time I am cloning, cmaking, and applying the patch BEFORE making the first time
19:34 GreenDimond   Because it wasnt linking the executable
19:35 Krock         weird
19:39 GreenDimond   Krock: Now that the executable is linked..
19:39 GreenDimond   can confirm: works with your script
19:40 Krock         yay. I don't know why my PR fixed it.. but well *happy shruggy*
19:40 GreenDimond   can confirm: works with my mod
19:40 GreenDimond   \o/
19:40 GreenDimond   but now
19:40 GreenDimond   minetest.[de]compress :)
19:41 Krock         test resuults ~~> PR
19:41 GreenDimond   Yes yes
19:43 Krock         still investigating
19:43 GreenDimond   I am surprised the compression bug has never been found
19:43 GreenDimond   The feature has been around for a while
19:43 GreenDimond   I guess no one had the idea to take up less storage space 8)
19:45 Krock         compress & decompress work fine
19:45 Krock         it's the combination with the mod storage which is the issue
19:47 GreenDimond   oh noes
19:48 IhrFussel     Krock, so am I right? multicore CPUs cannot make use of that setting?
19:48 Krock         IhrFussel: check with htop but I don't think so
19:48 Krock         only multiple processes could make use of multiple cores
19:48 GreenDimond   I have an 8-core single thread e. cpu and the setting works for me
19:49 GreenDimond   at least I think it does
19:49 Krock         GreenDimond: so your CPU usage is at 100% entirely taken by Minetest?
19:49 GreenDimond   No
19:49 GreenDimond   Minetest only takes one thread
19:49 GreenDimond   its 12.5% used
19:50 Krock         did you mean: *one core?
19:50 GreenDimond   no
19:51 GreenDimond   8 cores
19:51 GreenDimond   one thread each
19:52 GreenDimond   AMD FX-8320, in case you're curious
19:53 GreenDimond   (yes, it's somewhat old)
19:53 IhrFussel     What I can see is that there are multiple emerge threads in htop and when I fly with fast through ungenerated underground pretty much all of them increase their CPU usage to 20-30%
19:53 * Krock       laughts in Core 2 Quad Q6700
19:54 Krock         8k AMD vs 3.3k Intel
19:54 * Fixer       i5
19:54 Krock         (cpu benchmark)
19:55 GreenDimond   Just wait til I get a threadripper >.>
19:56 * Jordach     laughs in Haswell Refresh
19:56 Jordach       soon to be Coffee Lake:tm:
19:58 GreenDimond   Coffee Lake vs Threadripper 2990WX
19:58 Jordach       smells like a waste of power to me
19:58 GreenDimond   Nah dude
19:58 Jordach       >plays MT on threadripper
19:58 GreenDimond   Blender performance is gonna be instant
19:58 Jordach       >uses 1/64 threads
19:59 GreenDimond   I dont need anything fast for MT
19:59 GreenDimond   its Blender I care about
20:00 Jordach       <GreenDimond> I dont need anything fast for MT
20:00 * GreenDimond slaps AMD Threadripper .. this bad boy can render so many UHD scenes
20:00 Jordach       so you don't play on VanessaE's servers then
20:00 GreenDimond   fine
20:00 GreenDimond   I dont need anything fast_er_ than what I have
20:01 GreenDimond   Dreambuilder runs like thick butter on my current rig
20:02 GreenDimond   Not smooth butter
20:02 GreenDimond   but butter
20:02 Jordach       i've seen actual crude oil flow better than that :^)
20:03 Krock         GreenDimond: de/serialization tests: https://pastebin.com/raw/4qDvhGV1
20:03 Krock         it really fails only when reading from the file
20:04 Krock         set_string works fine with get_string afterwards
20:04 Krock         I'm now trying to remove minetest.compress from this mess
20:04 nerzhul       Krock what is the problem ?
20:04 Krock         nerzhul: mod storage is not persistent
20:05 nerzhul       wtf
20:05 Krock         if you write binary data to it, there's a data loss
20:05 nerzhul       it's stored in the database
20:05 GreenDimond   \o/ yay big bugs
20:05 nerzhul       data loss ? problem in the serialization then ?
20:05 nerzhul       or sqlite ?
20:05 Krock         somewhere around sqlite for sure
20:05 Krock         either read or write, not sure which one
20:05 nerzhul       did you tried with postgres ?
20:06 Krock         moment
20:06 GreenDimond   Is this considered a blocker? :v
20:06 Krock         GreenDimond: no. not a critical issue
20:06 nerzhul       i think it should be
20:06 nerzhul       and we should release at a point too :p
20:06 GreenDimond   Darn
20:06 nerzhul       maybe i will take time to release an android apk with one subset of the fixes
20:06 GreenDimond   I hope its fixed before 5.0 tho
20:06 Krock         nerzhul: what. there's no database for it. it's just files
20:06 nerzhul       if it's blocker it's the goal :p
20:07 Krock         mod storage has always been a file afaik
20:07 GreenDimond   It's always been plain text
20:07 nerzhul       mod storage API ? no
20:07 GreenDimond   Yes
20:07 nerzhul       it's in database, file is not a possible backend
20:07 GreenDimond   modstorage is in a plain text file
20:08 nerzhul       i developed it
20:08 GreenDimond   *shrug*
20:08 GreenDimond   its plain text
20:08 nerzhul       are you sure you are using the mod storage api ?
20:08 Krock         nerzhul: ikr but could it be that you slept? there's a mod_storage directory inside the world
20:08 Krock         local m = minetest.get_mod_storage()
20:08 Krock         see my script above
20:08 nerzhul       it's not a folder it should be an object to store data
20:08 Krock         it's portable
20:09 nerzhul       maybe a legacy mode ?
20:09 nerzhul       oh yes maybe
20:09 nerzhul       then player must migrate to sqlite or better postgres :)
20:09 Krock         nerzhul: how's the setting called to point to another new backend?
20:10 GreenDimond   Make sqlite default please
20:10 GreenDimond   players installing through things like ppa arent going to be able to use sqlite if they have to change it manually
20:10 Krock         I cannot find anything inside settingtypes.txt or world_format.txt
20:11 GreenDimond   world.mt has a database setting
20:11 GreenDimond   backend = sqlite3
20:11 Krock         that's for the map
20:11 nerzhul       or it's a metadataref
20:11 nerzhul       MetaDataRef
20:11 nerzhul       it was refactored by ruben i remember now
20:11 nerzhul       it's why there is plain files
20:12 GreenDimond   metadataref ~= modstorage...
20:12 Krock         GreenDimond: it is
20:12 Krock         modstorage inherits from metadata
20:12 Krock         same as the player meta
20:12 GreenDimond   oh
20:12 GreenDimond   til then
20:12 Krock         was about to say that lol
20:13 GreenDimond   So now we have established that it is indeed plain text, nerzhul, but now can we discuss fixing it?
20:13 nerzhul       it's not nice
20:13 nerzhul       erf
20:13 nerzhul       we should save that to the world db :(
20:14 nerzhul       fix the plain if possible Krock, we can try to add db backend in 5.1
20:14 Krock         still investigating where the issue exactly is
20:14 GreenDimond   I dunno, saving to a separate db would be nice because it can be easily deleted like the current state
20:15 nerzhul       GreenDimond not needed a single db table is sufficient
20:15 Krock         would also allow async actions
20:15 nerzhul       and permits to keep coherence with world
20:15 GreenDimond   I know its sufficient but is it convenient?
20:15 GreenDimond   How am I supposed to clear modstorage
20:15 GreenDimond   If its all in the world db
20:15 nerzhul       truncate table mod_storage;
20:15 nerzhul       ?
20:15 GreenDimond   I suppose
20:16 GreenDimond   I just found deleting it easier
20:16 Krock         drop table mod_storage would work too, due to legacy which will be added
20:16 nerzhul       yep but we should keep consistence in world if we use db :) and db offer it
20:17 GreenDimond   Hopefully the plain text method can be fixed, anyway.
20:22 Krock         after the first number >= 0x80 mod storage fails
20:22 nerzhul       string instead of bytes ?
20:23 Krock         the saved data is also 3 bytes longer
20:23 Krock         dunno
20:24 Krock         nvm, longer by 19 bytes actually
20:29 Krock         \ufffd everywhere
20:32 Krock         json[it->first] = it->second;   the issue must be around here
20:32 Krock         saving is broken
20:33 GreenDimond   Is that why it works in the same minetest session but not after restarting?
20:34 Krock         https://github.com/open-source-parsers/jsoncpp/issues/724
20:34 Krock         GreenDimond: yes, same as I mentioned minutes ago :D
20:35 GreenDimond   :]
20:35 Krock         hence "somewhere around loading and saving"
20:35 GreenDimond   Don't mind me - just pointing out the obvious ;D
20:36 Krock         k sir :)
20:37 GreenDimond   Krock: erm
20:37 GreenDimond   I think its loading
20:37 GreenDimond   not saving
20:37 GreenDimond   If I copy the data in the modstorage file before exiting, and paste after exiting, there is no change.
20:37 Krock         how come? saving inserts many more characters
20:38 Krock         trying to read the stuff as c strign
20:39 GreenDimond   maybe it is saving
20:39 GreenDimond   idk
20:39 GreenDimond   ill leave it to you, I dont know what im doing
20:43 Krock         GreenDimond: for now, consider converting your binary data to base64 and back
20:43 Krock         it entirely misses the point, thoguh
20:44 GreenDimond   wait what
20:45 GreenDimond   if that doesnt compress the data any more than serialize I wont bother
20:45 GreenDimond   is there even a method provided to do that?
20:46 paramat       IhrFussel about number of emerge threads, docs may be misleading, see https://github.com/minetest/minetest/issues/4879#issue-194766354
20:47 Krock         GreenDimond: the idea would basically be to convert the compressed data to base64 before passing it to minetest.serialize. However, you can also omit latter and only save a single string
20:47 paramat       if there is anything running in 'on generated' (there usually is) it may not be faster because lua isn't threaded
20:47 paramat       those docs should probably be updated to remove the claim
20:48 GreenDimond   Krock: So I would do storage:set_string(id, base64(minetest.compress(minetest.serialize(data_table)))) where base64() converts it to base64?
20:49 Krock         GreenDimond: oh lol. UTF-8 is built on top of ASCII, using the 8th bit to determine whether it's UTF-8 or not
20:49 Krock         minetest.encode_base64(compressed_data)
20:49 IhrFussel     paramat, so I should just remove that line from my settings and choose the default?
20:49 GreenDimond   oh thats a thing
20:49 GreenDimond   sweet
20:50 GreenDimond   will try
20:50 Krock         minetest.serialize is not needed if you don't want a table from it
20:50 GreenDimond   well the data is a table
20:50 GreenDimond   which needs to become a string
20:50 Krock         oh okay
20:50 GreenDimond   cool, hopefully this works. afk for a few hours now though.
20:51 Krock         so what JSON actually does is to correct the faulty UTF-8 encoding while writing to the file
20:51 Krock         except that we don't want that
20:52 Krock         GreenDimond: if you want to open an issue, then is https://github.com/open-source-parsers/jsoncpp the right place
20:52 Krock         not our problem (TM)
20:54 Krock         NEXT CUSTOMER
20:54 paramat       Ihr i'm not sure, not expert on how it works, maybe rogier is wrong
20:54 Krock         *makes beep noises*
20:58 IhrFussel     I will remove it for the next restart and check the lag values again...then we should have a definite answer
20:58 IhrFussel     Also is there any downside to increasing the unload_unused_data_timeout server side? I mean making it really high like 10 minutes?
20:59 IhrFussel     It uses more RAM of course but is there any other potnetial issue?
20:59 IhrFussel     potential*
21:02 IhrFussel     I have too much RAM so that will not be a problem... seriously like who needs 24 GB DDR4 RAM for a MINETEST server
21:02 paramat       the default is a clue to the optimum values
21:03 IhrFussel     But if the server keeps more mapblocks in-memory it should in theory reduce server lag a bit in case a player visits an area that was already loaded from file
21:04 Krock         aren't 2 GiB already enough for a regular server?
21:04 IhrFussel     And if the server gets to keep all mapblocks from several minutes ago it should mean less work...for example when players visit a major area
21:04 Krock         (RAM)
21:06 IhrFussel     My server uses 2.2 GB right now and my unused data timeout is set to 90 I think
21:06 IhrFussel     I just wanted to know if the chance of map file corruption could increase if I set that timeout too high
21:07 IhrFussel     But I guess not cause it doesn't keep modified mapblocks in RAM correct?
21:07 Krock         it keeps all active mapblocks in RAM
21:08 IhrFussel     But I mean it doesn't wait until the timeout is reached before it saves modified blocks
21:08 Krock         modified ones are saved each save interval (setting) and unloaded after your 90s
21:08 Krock         whereas it saves again on unloading if necessary
21:08 Krock         (dirty flag)
21:09 IhrFussel     So setting it high shouldn't make a big difference cause all iportant changes are saved every 15 secs
21:09 paramat       29s seems rather short and random
21:10 IhrFussel     29 seconds is like nothing... if a player visits a big area with lots of meta and leaves and another visits the same area fter 1 minute the server doesn't have it anymore in RAM and needs to load it from DB again
21:10 IhrFussel     If I understand the mechanism correctly
21:11 IhrFussel     I mean if the area is protected then there is also no modification so the mapblock is read-only right?
21:11 Krock         IhrFussel: on the other side node timers and ABMs will keep on modifying the mapblock
21:11 Krock         causing save events
21:11 paramat       i feel like the default should possibly be raised
21:12 IhrFussel     Krock, but... does it only save certain mapblocks then or all in RAM?
21:12 nerzhul       all modified mapblocks are saved on each mapblock_save_interval
21:12 nerzhul       not all, just modified
21:13 IhrFussel     So a very high in-memory value should not make it worse
21:16 IhrFussel     Wait...it could slow down the server if too many in-memory mapblocks contain nodetimers
21:18 nerzhul       the slowdown is generally due to many lua callback, timers, abm et
21:19 nerzhul       etc
21:20 Krock         IhrFussel: only those which are modified by set_node, vmanip or players
21:21 IhrFussel     But you said the timers run on all in-memory mapblocks... now imagine a few thousand need to be checked
21:22 Krock         oh right. node timers are saved in the mapblock data too
21:22 Krock         yes, in that case it'll also affect those with triggered node timers
21:29 aerozoic      apparently i dont understand registering an alias...
21:30 aerozoic      I want to tame a penguin using fish not defined in the penguin code
21:30 aerozoic      penguins require "ethereal:fish_raw" but i don't have eathereal. I do have "fishing:fish_raw" so how do i make that work?
21:32 aerozoic      i tried:  minetest.register_alias("fishing:fish_raw", "ethereal:fish_raw")
21:32 aerozoic      but that gives an error
21:32 IhrFussel     aerozoic, Mobs Redo has its own alias function..one sec
21:33 IhrFussel     mobs:alias_mob(oldname,newname) ... not sure if that will work in your case
21:34 aerozoic      no, the fish itself is not a mob, just a craftitem.
21:36 IhrFussel     aerozoic, what's the error register_alias() gives?
21:36 aerozoic      2018-12-28 16:12:50: WARNING[Main]: Not registering alias, item with same name is already defined: fishing:fish_raw -> ethereal:fish_raw
21:38 IhrFussel     I'd say it means you have to remove the item that's defined as etheral:...
21:38 IhrFussel     It is defined somewhere it seems
21:39 aerozoic      ok i think i figured it out, the dev page describing the function is backwards.
21:40 aerozoic      switching the two items made it work
21:41 aerozoic      this doesn't solve my problem though :(
21:42 aerozoic      maybe i need the override function
21:45 IhrFussel     aerozoic, maybe you have to add your own fish item to the actual definition of the mob
21:45 aerozoic      Yes that is a last resort, i don't want to modify mobs code
21:46 IhrFussel     Well it's only 1 line in penguin.lua
21:47 aerozoic      mobs is actively updated on a regular basis. Preserving the edit would require special treatment in order to stay up to date.
21:48 aerozoic      Surely there is a way
21:54 aerozoic      Is there a way to give an item more than one name?
21:55 IhrFussel     aerozoic, the "engine" is updated regularly but the animals themselves are barely touched
21:57 IhrFussel     Almost everything happens in mobs/api.lua
21:58 aerozoic      If i understand correctly, i can use minetest.override to modify the penguin :)
22:14 aerozoic      nope :(
23:02 rdococ        pffgfghfghfgt
23:29 benrob0329    we are not doing sandstone justice https://cdn.pixabay.com/photo/2014/09/23/11/29/antelope-canyon-457495_960_720.jpg
23:47 rdococ        hmm
23:47 rdococ        wmore colors of sandstone and a different texture for it