Time  Nick         Message
00:07 Megaf_       and after months without a single crash I finally get a crash http://paste.debian.net/94758/
00:08 Megaf_       What could have cause that?
00:09 Megaf_       sapier, It's something to do with connections.cpp
00:13 troller      its sapier asserts on every second line
00:13 troller      recieve wrong packet => server crash
00:15 troller      34 asserts in connection.cpp
00:15 troller      enterprise quality code!
00:16 Megaf_       troller, can you fix?
00:16 troller      in freeminer only, if you show this line
00:16 iqualfragile false
00:17 troller      maybe it was already fixed
00:17 Megaf_       nope, just did a git pull
00:17 Megaf_       nothing related to connection.cpp
00:18 iqualfragile Megaf_: best idea would be to file a bug or wait 8-9 hours
00:18 iqualfragile and then tell sapier when he is online
00:18 iqualfragile oh wait he is
00:18 iqualfragile strange, usually goes offline a lot earlier
00:18 Megaf_       iqualfragile, usually he is really responsive and helpful when I have a problem
00:19 iqualfragile well, unless you ask him to make something in mobf usable, in that case you have to wait half a year
00:43 sapier1      END_DEBUG_EXCEPTION_HANDLER is a old minetest debug define ... megaf didn't you have performance issues? why do you have ancient debug code enabled?
00:45 sapier1      if you tell minetest to crash on any non critical error it will do this ... but you shouldn't complain about it doing what you request it to do
00:49 celeron55    those handlers are enabled on release builds because they usually give more useful output than nothing at all
00:49 celeron55    use CMAKE_BUILD_TYPE=Debug to make it not catch them and use a debugger
00:51 sapier1      what?
00:52 sapier1      we use exceptions for passing return values in one part of minetest but crash in others? ... ok guess I need to check if it'd be best to remove that define there
00:56 sapier1      the sad thing about this handling is it doesn't seem to provide any usefull information
00:58 iqualfragile sapier1: i guess packets can allways be damaged
00:59 sapier1      that's why invalid data exeptions are cought prior this define
01:00 sapier1      the log doesn't seem to tell what exception happened
01:34 Megaf_       It's not my fault that there ancient debug code, I have chosen to use Release Build type and I still get that thing...
02:08 Megaf_       Folks, how does this affects performance?
02:08 Megaf_       # Maximum number of forceloaded blocks
02:08 Megaf_       max_forceloaded_blocks = 1
04:06 hmmmm        oh noes :(
04:06 hmmmm        I just realized something that messes up my biome plan
04:07 hmmmm        at the edge between, say, a flatland and a flat highland, there will be a mini-biome riding along the edge which would be the craziest mountain you can find
04:08 hmmmm        so this kind of terrain http://i.imgur.com/rwcZ9Pc.png  would generate this kind of variance map  http://i.imgur.com/Tq240cN.png
04:17 RealBadAngel hi
04:18 hmmmm        also my fast gaussian blur is a mere, inaccurate approximation of the mean in a certain radius.  the real thing:  http://i.imgur.com/JIds934.png    fast approximation: http://i.imgur.com/NGrrkl3.png
04:19 RealBadAngel hmmmm, ive coded lately gausian blur for shaders and normal maps, wanna see that code?
04:19 hmmmm        if i use the fast approximation for variance calculation, it'll get below 0 which variance is theoretically not supposed to do
04:19 hmmmm        RealBadAngel, no thanks
04:23 RealBadAngel btw, lookin at those 2 pictures, to get effect shown on the 1st one propably you should use bigger kernel
04:23 RealBadAngel maybe 9x9
04:28 RealBadAngel about #1233 , can we have this merged? it has been remade as c55 wanted
04:28 ShadowBot    RealBadAngel: Error: Delemiter not found.
04:29 RealBadAngel huh? https://github.com/minetest/minetest/pull/1233
04:30 hmmmm        dunno, that's a question for c55
04:31 RealBadAngel https://github.com/minetest/minetest/issues/486
04:32 RealBadAngel he already answered, so i assume he votes yes
04:36 hmmmm        looks good, but personally I'd break up Sky::render() into multiple smaller functions that do simpler tasks
04:38 RealBadAngel maybe in the future, atm this was lying on shelf way too long
04:39 RealBadAngel i do have plans for skybox but i wont start coding that anytime soon
04:39 RealBadAngel i am busy with further shaders coding
04:40 RealBadAngel take a look how united nodes shader looks like: https://github.com/RealBadAngel/minetest/blob/shaders/client/shaders/nodes_shader/opengl_vertex.glsl
04:43 RealBadAngel im done with one single shader for all the nodes, selectable by tile, not whole node. so water surface shaders are possible now. also ive moved generation of shaders away from runtime
04:43 RealBadAngel the last has huge impact on performance
04:46 RealBadAngel next on the desk is allow multiple rendering passes and shaders postprocessing
04:46 RealBadAngel when done water will have reflections
04:47 RealBadAngel also effects like lens flare, god rays etc will become easy to add
04:47 kaeza        suddenly: render to texture
04:47 * kaeza      dreams
04:48 RealBadAngel nothing easier
04:48 RealBadAngel i do have forked irrlicht app that does it
04:48 RealBadAngel https://github.com/RealBadAngel/terrainTest
04:49 RealBadAngel all whats needed is here
04:53 RealBadAngel kaeza, btw if you have linux working again you could check that branch and see if you can still reproduce fog issue
04:54 kaeza        14.04 borked support for my modem
04:54 RealBadAngel lol
04:55 kaeza        I've set up MinGW but can't build MT :(
04:55 kaeza        does not happen on Win anyway
04:55 kaeza        sorry :/
04:56 RealBadAngel anyway im pretty sure that this wasnt caused by shaders code
04:56 RealBadAngel weather was also touchin the fog
05:24 Megaf_       Night all
12:39 iqualfragile sfan5: provided the backtrace
13:07 BlockMen     sapier, #1240 fixes #1223 for me.
13:07 BlockMen     Side notice: It seems there is no red background for death screen anymore
13:07 ShadowBot    https://github.com/minetest/minetest/issues/1240 -- Fix broken local formspec handling by sapier
13:07 ShadowBot    https://github.com/minetest/minetest/issues/1223 -- MT (sometimes) crashes when leaving pause menu under Windows
13:44 sapier       yes blockmen, for what I read red background wasn't there in all situations anyway
13:46 sapier       correct me if that's wrong, if we wanna have a red screen on death that shouldn't be a big deal
14:14 BlockMen     sapier,
14:14 BlockMen     bg[#oldredcolor;true] ;)
14:15 sapier       bg color can be set by formspec?
14:15 sapier       but deathscreen wasn't read on original code for any of my tests
14:16 sapier       can you check that inventorys and formspec work as expected. seems to work for me on linux but I'm still working on the android merge, for some reason I get invalid refcounts there
14:19 BlockMen     sapier, added > half year ago https://github.com/minetest/minetest/blob/master/doc/lua_api.txt#L930
14:19 BlockMen     and its bgcolor, sry
14:19 BlockMen     old color: https://github.com/minetest/minetest/blob/eda9214f81b32d9606f425e3777616e8cfc0a44f/src/guiDeathScreen.cpp#L121
14:20 BlockMen     and i can do later ;)
14:20 sapier       are you sure this did work at all?
14:22 sapier       well actually this doesn't matter I'll add it anyway ... once I found out why my andorid port doesn't behave as expected
14:24 BlockMen     worked what? #1240? yea
14:24 ShadowBot    https://github.com/minetest/minetest/issues/1240 -- Fix broken local formspec handling by sapier
14:24 sapier       I was talking about red death screen
14:26 sapier       on android I've currently a problem someone is grabing formspecs and doesn't release them ... no idea why there's a difference to x86
14:28 BlockMen     yes, old deathscren hat red for sure: http://i.imgur.com/KAVfnRZ.png
14:28 * BlockMen   has no idea on andriod
14:28 BlockMen     -on +of
16:38 sapier       can someone else please test #1240 too? I'd prefere not to have to fix it again tomorrow because of missing some other case ;-)
16:39 ShadowBot    sapier: Error: Delemiter not found.
16:41 proller      why you use assert in every second line ?
16:41 sapier       to show errors that need to be fixed instead of hiding them
16:42 proller      and to crush all clients and servers at any small error?
16:42 sapier       assert is supposed to be a nop in release builds
16:43 sapier       not getting memory or having completely messed up data isn't a small error
16:43 proller      most users/servers use -dev version
16:43 sapier       to be more precise messed up INTERNAL data
16:44 sapier       then those users server owners do wrong thing, dev builds are supposed to show up broken code not hide it
16:46 sapier       asserts are to be placed for things that can't happen ... of course the decision what can happen may be wrong, but in any case if a dev assumes something not to happen he's most likely not gonna handle that case correct so even in this case the assert is right
16:46 proller      stop stop stop
16:46 sapier       I guess you're talking about megafs problem right?
16:47 sapier       invalid incoming data is handled prior the catch all he ran into so this has to be something different
16:48 sapier       problem is that those traces do only show that some exception occured but not what exception
16:48 celeron55    sapier: minetest doesn't turn asserts into nops
16:48 celeron55    in release builds
16:48 sapier       it doesn't?
16:48 celeron55    no; it's intentional and has been so right from the start
16:48 sapier       then how's the official way to do asserts?
16:48 * proller    facepalm
16:49 proller      34 asserts in connection
16:49 proller      most of them happen
16:49 celeron55    sapier: if you want such an assert, implement it in some way
16:49 proller      celeron55, review connection.cpp please
16:49 celeron55    most asserts in minetest do stuff like check that a pointer is not NULL
16:49 sapier       do you really suggest any file to implement it's own assert?
16:49 celeron55    in which case it doesn't make any sense to let execution pass through in release builds
16:49 celeron55    because it would just explode even worse
16:50 celeron55    sapier: no; add something like debug_assert() in debug.h for the purpose of debug-build-only asserts
16:50 celeron55    then you can use them happily ever after
16:51 sapier       well assert handling I know is are always used that way still they're nops in release build
16:51 celeron55    there are two schools of assert handling in this world
16:51 celeron55    the other one is this
16:52 sapier       proller if you didn't do a wc only you'd have seen most if not all asserts in connection.cpp are no continue asserts
16:53 sapier       e.g. if reliable packet handler  reads "non-reliable" from a packet that has been identified to be a reliable packet before
16:55 celeron55    anyway, if what proller is saying is true, you should fix it
16:55 celeron55    it shouldn't be possible to crash a server or a client by sending it data over network
16:55 sapier       if those exceptions would happen he'd be right
16:56 proller      its true 8(
16:56 proller      i catch some on my servers
16:56 celeron55    the right thing to do in case of invalid data received is to ignore the packet, and possibly disconnect
16:56 sapier       provide error logs and I'm gonna fix it
16:56 proller      remove all asserts - good fix
16:57 celeron55    that sounds like a very bad idea
16:57 sapier       that's how it's supposed to be done, most asserts are in later packet processing where packet is supposed to have already be checked to be of some type. In case it isn't something was wrong about the way it's been processed.
16:57 celeron55    i.e. it sounds like buffer overflow and other kinds of vulnerabilities
16:58 sapier       e.g. if we try to send out a sequence number out of sequence number range
16:58 sapier       or if someone tries to send a packet at a non existent channel
16:59 sapier       or someone provides a invalid packet to be sent. I do not tolerate our own code to try to send broken packets.
17:00 sapier       proller if you can provide logs I can fix em if not I can't
17:01 proller      i have 5.5G logs, ready to read ?
17:01 sapier       nope I don't accept manual dos attacks
17:01 sapier       https://github.com/sapier/minetest/commits/android_merge_branch_5 rebased android port
17:05 sapier       you can try grep assert first if you don't wanna do it manually, hopefully the line numbers tell what's happening
17:09 Megaf_       sapier, I will test it
17:09 sapier       what do you want to test megaf? by now I don't have an idea how to catch your issue
17:09 RealBadAngel celeron55, what do you think of #1233 ? can we have it merged?
17:10 ShadowBot    RealBadAngel: Error: Delemiter not found.
17:10 sapier       did anyone else have an idea?
17:10 RealBadAngel https://github.com/minetest/minetest/pull/1233
17:10 Megaf_       sapier, I'm going to be a test of ANdroid build
17:10 Megaf_       Android*
17:10 Megaf_       sapier, That workarround you did works perfect
17:11 sapier       oh :) good by now I don't think anyone did build itself by now
17:11 Megaf_       Now I get no timeouts or disconnections even when CPU load is more than 100% (Yesterday a friend of mine did a LUA loop on lua controllers)
17:12 sapier       hmm I wonder why arm is that bad on log handling
17:12 Megaf_       ARM is a much simpler CPU, good to do less things at the same time
17:13 Megaf_       but it does those thins very quickly
17:13 RealBadAngel android can run not only on arm cpus
17:13 sapier       still it somehow feels wrong to be fixed that simple
17:13 RealBadAngel i do have android_x86 installed
17:13 Megaf_       RealBadAngel, we are talking about my server that runs on ARM
17:13 sapier       I don't really believe it to be fixed completely
17:13 Megaf_       sapier, as I said, is a work around
17:14 Megaf_       that works and has no side effects
17:14 Megaf_       affects
17:14 sapier       ok ok I guess I should be more precise when talking about my android/arm port :-)
17:14 Megaf_       sapier, now, about Android port, Android does run on Intel, my Android phone is an Intel Atom powered
17:15 Megaf_       also, I run Android x86 on my Netbook
17:16 sapier       well you can try it but I doubt it's gonna build there
17:16 celeron55    RealBadAngel: so it behaves exactly the same as currently if there are no textures in the texture pack?
17:16 RealBadAngel celeron55, exactly
17:16 celeron55    are the texture names future-proof
17:17 celeron55    they seem rather simple but... i guess any future functionality will just have more complicated names
17:18 RealBadAngel i can see no reason why sun.png for sun should have any other name ;)
17:18 celeron55    the code is a bit messy
17:18 RealBadAngel why do you think so?
17:19 celeron55    a function for drawing each thing would make the selection logic easier to see
17:19 Calinou      we should really have an invisible.png by the way...
17:19 celeron55    now it's spread around between multi-ten-line drawing code
17:20 sapier       for(u32 i=0; i<4; i++) << I know this is spread around code but is it really necessary to force index variables to be 32 bit?
17:21 RealBadAngel for such thing i would introduce new variables or something then. by now texture pointerr are such flags (get_texture returns 0 if texture is not found)
17:21 RealBadAngel so i used that for logic
17:23 celeron55    well, if nobody finds a problem in it soon, i guess it can be merged
17:24 RealBadAngel vanessaE have it tested already
17:25 RealBadAngel also i made it such way that tinting works only if tonemap is provided, so both ways are possible
17:25 RealBadAngel well, 3 ways, including old one
17:25 sapier       :-) you'd not match ShadowNinja's coding style requirements :-) but that's not matching is quite simple
17:26 VanessaE_    yeah, I've tried both tinted and non-tinted modes (the latter by accident), they work fine and as expected
17:26 sapier       I don't see any obvious bugs in there too rba imho 1233 can be merged
17:26 RealBadAngel celeron, could you also take a look at this? https://github.com/RealBadAngel/minetest/commit/03936cf4a0e20ce9202b6a8bbcdd603661dc3e86
17:27 sapier       float offset=(1.0-fabs(sin((m_time_of_day - 0.5)*irr::core::PI)))*511; black magic?
17:27 RealBadAngel ;) yeah, somethin like that ;)
17:28 sapier       maybe add a comment what you're calculating there, could help if you wanna check it in half a year
17:28 RealBadAngel split day in half basically
17:28 VanessaE_    normalized to the X dimension of a tonemap, which is 512 px
17:28 RealBadAngel to get noon/midnight
17:29 sapier       well then this is the comment because I wouldn't have had any chance to read that from that line ;-)
17:29 RealBadAngel ok, i will put some comments there
17:30 sapier       thanks
17:32 celeron55    RealBadAngel: i'm not going to try to review that; i just hope you are around to fix it when it turns out it doesn't work for people
17:32 celeron55    (which means, don't let it be merged when you aren't available for the next days)
17:32 RealBadAngel celeron55, shaders are workin now for more ppl than before
17:34 sapier       *smile* don't bother rba, it's quite normal to do mistakes, just be there to fix them
17:34 RealBadAngel and bugs... bugs happens always ;) idk why that glowing water passed into commit
17:34 RealBadAngel not only me was testing before merging
17:35 sapier       I did even like the glowing water, could be a nice to have feature ;-) ... not for all water but maybe for some special sort of water
17:35 RealBadAngel hehehe
17:35 celeron55    well you can get it if you set water to be a light source
17:35 celeron55    or at least it's a bug if that doesn't happen
17:36 RealBadAngel lava glows, so it works i think
17:36 sapier       well I wouldn't even have had the idea without this bug ;-)
17:37 RealBadAngel btw, the commit with shaders unite allows completely new things. shaders can be set by tile, not the node, this means water surface shaders possible
17:38 RealBadAngel or lava surface shaders
17:39 PenguinDad   RealBadAngel: I've noticed one little problem with your shaders branch. water shaders fail to compile when waving water is disabled
17:40 RealBadAngel theres no water shader anymore
17:41 RealBadAngel youre talking about this branch? https://github.com/RealBadAngel/minetest/tree/shaders
17:42 PenguinDad   RealBadAngel: yes and this the error message http://sprunge.us/TYQj
17:42 PenguinDad   +is
17:42 RealBadAngel lemme check the reason
17:45 RealBadAngel ok, thats && ENABLE_WAVING_WATER missing
17:45 RealBadAngel will push fix in a moment
17:50 RealBadAngel PenguinDad, fixed
17:52 RealBadAngel PenguinDad, how does the shaders from that branch work for you? can you see any change in performance?
17:53 PenguinDad   RealBadAngel: there's no real perfomance change for me with my GTX 650
17:55 RealBadAngel with my old radeon i have some fps more than before
17:55 RealBadAngel i guess when card is fast enough for old ones, you wont notice any gain
17:56 VanessaE_    it also depends on the user's view range limit
17:56 VanessaE_    if they're already hitting their fps cap with their current view range cap, they won't see any improvement.
17:58 RealBadAngel but one thing everybody will notice, mesh updates are way faster now. and there wont be such glitch like before when world has loaded, textures shown and after a few seconds shaders got loaded and applied
17:59 RealBadAngel that was a very bad idea to let shaders compile on mesh updates
18:29 Megaf        folks, not long ago the log used to show the player's IP address, now it does not shows it anymore, can you please revert that?
18:30 Megaf        15:28:41: ACTION[ServerThread]: MegafSever joins game.
18:30 Megaf        15:28:41: ACTION[ServerThread]: MegafSever joins game. List of players: Megaf RexVlogs10 gold
18:33 PilzAdam     Megaf, sapier did it in https://github.com/minetest/minetest/commit/142e2d3b74ad886eed83b0fc9d6cfea100dae10a
18:33 PilzAdam     sapier, fix it!
18:34 Megaf        :)
18:39 RealBadAngel sapier, https://github.com/RealBadAngel/minetest/commit/b899f11cf40a309af4959b5632003572f2bd8e7c
18:39 RealBadAngel enough?
18:40 sapier       fine for me yes
18:43 RealBadAngel ok, squashed. can somebody merge it?
18:43 sfan5        can't you do it yourself?
18:44 RealBadAngel i can push it directly from my hdd
18:45 RealBadAngel so issue will stay open
18:50 sapier       PilzAdam ... there's a small word called "please" ;-)
18:50 sapier       it's quite helpfull sometimes
18:50 RealBadAngel i mean the commit will be different than that from opened issue
18:51 sapier       just add a comment there about it beeing fixed by the other one
18:52 sapier       are you about to merge right now rba or can I push the ip fix first?
18:53 RealBadAngel you can push your fix first, travis is thinkin about mine anyway
18:54 RealBadAngel well, travis says good to go
18:54 sapier       then push yours first
18:54 RealBadAngel but how can i push it without that green merge button?
18:54 RealBadAngel the one i was told not to use ;)
18:55 sapier       exacly ;-) don't use the button but do it on command line ;-)
18:55 RealBadAngel how then?
18:55 kaeza        git checkout master; git rebase your_branch; git push
18:55 sapier       megaf what log syntax do you expect? "playername [ip] joins game."
18:55 sapier       ?
18:56 RealBadAngel lets see
18:56 Megaf        sapier, Yes, that would be great, and also, if it's possible, some info about players version
18:57 Megaf        like. [date - time] PlayerName [234.456.123.456] Minetest 0.78 34235 - Joins game
18:57 VanessaE_    knowing the client version would be useful but I doubt that's possible with the current protocol
18:57 RealBadAngel but hold on, my branch is a branch of forked minetest, not the original one
18:59 sapier       client info would be a lot of additional code there, code that shouldn't be at that location. I suggest doing this by a mod
19:00 kaeza        RealBadAngel, so you are working on `master' branch?
19:00 RealBadAngel minetest (original one) and my fork are separate git repos on my hdd
19:00 RealBadAngel i am workin on forked repo, not the original (master) one
19:01 RealBadAngel i can commit to master by copying commit files to master repo and pushing it
19:02 kaeza        you should learn how to use branches; it makes these things easier
19:03 RealBadAngel so i do have branches, but i dont know how to push commits made by others (including me) to master repo
19:03 RealBadAngel other than pushing them directly
19:03 RealBadAngel and using that damn green button
19:04 PenguinDad   RealBadAngel: remote branches
19:04 RealBadAngel just tell me anybody. theres commit made by me, in my own fork. what are you doing to get it merged....
19:09 RealBadAngel how one can merge any commit from an issue?
19:10 RealBadAngel other than picking diff and applying it to master...
19:10 PilzAdam     RealBadAngel, go locally to your master branch, be sure that its in sync with upstream/master; then just cherry-pick the commits you want and push
19:12 RealBadAngel "just cherry-pick" - how?
19:12 PilzAdam     git cherry-pick ef8600bf0e43a59b2abfc4fd4875332d944c2daf
19:13 PilzAdam     ^ that would be your sun and moon texture commit
19:17 RealBadAngel http://i.imgur.com/owQ1Gvj.png
19:18 RealBadAngel my git client does have such commands but it doesnt show any commits waiting for merging
19:18 RealBadAngel it shows only already merged ones
19:18 PilzAdam     what about that giant "Cherry pick" button?
19:19 sapier       hmm funny anyone could tell about the command line commands but we fail for simple gui operations
19:19 RealBadAngel it can pick only from list you can see
19:19 RealBadAngel which are already merged commits
19:19 PilzAdam     just use the CLI then
19:20 VanessaE_    > RealBadAngel
19:20 VanessaE_    > CLI
19:20 VanessaE_    hah!
19:20 Megaf        sapier, ok then, but please, IP is really important
19:21 sapier       ip isn't a big deal as that information is handled separate
19:22 RealBadAngel sapier, that GUI doesnt show anything waiting to be merged... it can propably show commits if i create local branch in minetest repo
19:22 sapier       well I don't know that gui so I'll not be helpfull for it
19:23 RealBadAngel it serves well for maintainin my own repos, but idk how to use it to commit to mt
19:24 PilzAdam     hmm... "Local Branches (1)" can you somehow pull the branch you want to merge to your local repo?
19:24 RealBadAngel repos are separate
19:25 RealBadAngel so sun/moon branch is a branch of my own repo not the minetest one
19:25 sapier       rba I'm gonna merge it myself ;-)
19:25 RealBadAngel i already know how thx to PA
19:26 PilzAdam     why isn't the sun/moon branch listed in the GUI?
19:26 sapier       no need to do it ?
19:26 RealBadAngel because its minetest repo
19:26 RealBadAngel not my forked one
19:34 RealBadAngel PilzAdam, http://i.imgur.com/hjGqEot.png
19:34 RealBadAngel whats wrong with it?
19:34 PilzAdam     whats the output git branch?
19:35 RealBadAngel im in minetest repo
19:36 RealBadAngel master
19:37 PilzAdam     you first need to get the textureable_sun branch into your local repo
19:37 PilzAdam     you can do that with  git fetch https://github.com/RealBadAngel/minetest.git
19:37 PilzAdam     then the cherry-pick should work
19:38 RealBadAngel jeez, applying diff from commit is way easier...
19:38 PilzAdam     I guess your local repo setup is wrong
19:38 PilzAdam     thats why its so hard for you
19:39 RealBadAngel http://pastebin.com/gY21J9kh
19:39 RealBadAngel doesnt work
19:40 PilzAdam     I guess we better discuss this in PM
19:40 RealBadAngel ok
20:22 RealBadAngel sapier im done, you can push your things now
20:50 sapier       can anyone (exept BlockMen who already did) please check #1240? it shouldn't but may have sideeffects on anything handling formspecs e.g. inventory, node inventory, mod menus, pause menu, deathscreen, chat message. In case there's a n error in it it's most likely on those things directly beeing replaced by each other
20:50 ShadowBot    https://github.com/minetest/minetest/issues/1240 -- Fix broken local formspec handling by sapier
23:56 Megaf        20:56:39: ERROR[ServerThread]: ServerEnvironment::clearAllObjects(): Failed to emerge block (18,-66,11)
23:56 Megaf        This can't be good
23:57 Megaf        dozens of these lines each second again