| Time |
Nick |
Message |
| 01:04 |
|
Mantar joined #luanti |
| 01:09 |
|
simon816 joined #luanti |
| 01:26 |
|
FileX joined #luanti |
| 01:37 |
|
Kimapr_ joined #luanti |
| 02:36 |
|
aliasalreadytake joined #luanti |
| 02:54 |
|
simon816 joined #luanti |
| 03:06 |
|
v-rob joined #luanti |
| 03:14 |
|
SFENCE joined #luanti |
| 03:23 |
|
FileX joined #luanti |
| 04:02 |
|
MTDiscord joined #luanti |
| 04:04 |
|
simon816 joined #luanti |
| 04:22 |
|
sparky4 joined #luanti |
| 04:22 |
|
est31 joined #luanti |
| 04:23 |
|
sparky4 joined #luanti |
| 05:01 |
|
Kimapr_ joined #luanti |
| 05:24 |
|
simon816 joined #luanti |
| 05:47 |
|
Talkless joined #luanti |
| 05:52 |
|
aptgetefe joined #luanti |
| 05:52 |
|
aptgetefe joined #luanti |
| 05:54 |
|
kimapr joined #luanti |
| 06:02 |
|
Kimapr_ joined #luanti |
| 06:04 |
|
InFerNo_ joined #luanti |
| 06:09 |
|
gregon joined #luanti |
| 06:10 |
|
gregon left #luanti |
| 06:13 |
|
kimapr joined #luanti |
| 06:25 |
|
Kimapr_ joined #luanti |
| 06:28 |
|
Kimapr_ joined #luanti |
| 06:31 |
|
Glaedr joined #luanti |
| 06:36 |
|
sparky4 joined #luanti |
| 06:37 |
|
kimapr joined #luanti |
| 06:45 |
|
Kimapr_ joined #luanti |
| 06:48 |
|
sparky4 joined #luanti |
| 06:55 |
|
Kimapr_ joined #luanti |
| 06:59 |
|
simon816 joined #luanti |
| 07:02 |
|
kimapr joined #luanti |
| 07:12 |
|
kimapr joined #luanti |
| 07:18 |
|
Kimapr_ joined #luanti |
| 07:27 |
|
Kimapr_ joined #luanti |
| 07:33 |
|
luatic joined #luanti |
| 07:35 |
|
kimapr joined #luanti |
| 07:53 |
[MatrxMT] |
<birdlover32767> seems like keeping a list of all flooded nodes and checking for when they dry up is a solution |
| 07:53 |
[MatrxMT] |
<birdlover32767> then i can just use voxelmanip |
| 07:57 |
|
Kimapr_ joined #luanti |
| 08:07 |
|
kimapr joined #luanti |
| 08:07 |
|
serp6 joined #luanti |
| 08:11 |
|
sparky4 joined #luanti |
| 08:11 |
|
stephan48 joined #luanti |
| 08:13 |
|
sofar joined #luanti |
| 08:13 |
|
Kimapr_ joined #luanti |
| 08:23 |
|
Kimapr_ joined #luanti |
| 08:30 |
|
kimapr joined #luanti |
| 08:33 |
|
zleap joined #luanti |
| 08:33 |
|
zleap joined #luanti |
| 08:38 |
|
jaca122 joined #luanti |
| 08:43 |
|
sys4 joined #luanti |
| 08:44 |
|
est31 joined #luanti |
| 08:49 |
|
Pokey joined #luanti |
| 09:30 |
|
kimapr joined #luanti |
| 09:47 |
|
YuGiOhJCJ joined #luanti |
| 09:55 |
|
Kimapr_ joined #luanti |
| 10:04 |
|
Kimapr_ joined #luanti |
| 10:09 |
|
Guest62 joined #luanti |
| 10:12 |
|
kimapr joined #luanti |
| 10:16 |
|
mrkubax10 joined #luanti |
| 10:20 |
|
Kimapr_ joined #luanti |
| 10:36 |
|
simon816 joined #luanti |
| 10:40 |
|
Kimapr_ joined #luanti |
| 10:47 |
|
Kimapr_ joined #luanti |
| 10:54 |
|
kimapr joined #luanti |
| 11:01 |
|
FileX joined #luanti |
| 11:02 |
|
Kimapr_ joined #luanti |
| 11:09 |
|
Kimapr_ joined #luanti |
| 11:11 |
|
leobel13 joined #luanti |
| 11:18 |
|
Kimapr_ joined #luanti |
| 11:26 |
|
kimapr joined #luanti |
| 11:33 |
|
Kimapr_ joined #luanti |
| 11:39 |
|
ireallyhateirc joined #luanti |
| 11:39 |
|
Kimapr_ joined #luanti |
| 11:49 |
|
kimapr joined #luanti |
| 11:58 |
|
Kimapr_ joined #luanti |
| 12:06 |
|
kimapr joined #luanti |
| 12:07 |
|
Helenah joined #luanti |
| 12:12 |
|
Kimapr_ joined #luanti |
| 12:17 |
|
SliverFlowCipher joined #luanti |
| 12:20 |
|
kimapr joined #luanti |
| 12:27 |
|
kimapr joined #luanti |
| 12:42 |
bgstack15 |
I've been theorizing about hacking contentdb code so it would be possible to have a multi-backend ContentDB based on existing discussions on the repos. That is, a "contentdb proxy" that points to multiple real ContentDB instances. I would think it would be a stripped-down contentdb instance, that only implements lookups to the configured listed instances. |
| 12:44 |
bgstack15 |
So it wouldn't have its own users or packages, etc. I think the use case of using it merely from Luanti frontend is the only one it would work with correctly, so no logging in, voting (?), etc. Would such a ContentDB-proxy need to implement only api/endpoints.py (which contains all the /api/ endpoints)? |
| 12:46 |
MTDiscord |
<wsor4035> youll have to decide how to merge rankings, or handle sources without ranking for ordering |
| 12:46 |
|
simon816 joined #luanti |
| 12:46 |
MTDiscord |
<wsor4035> totally havent done anything similar to this |
| 12:46 |
MTDiscord |
<wsor4035> abandoned it as it was pretty pointless/useless |
| 12:47 |
MTDiscord |
<wsor4035> curious what your use case is? |
| 12:47 |
bgstack15 |
Yeah, this is just theorizing. We gotta start somewhere. |
| 12:47 |
bgstack15 |
Doing it so it is possible. |
| 12:47 |
bgstack15 |
I can't go to the moon, but maybe I can tinker on a contentdb proxy |
| 12:47 |
bgstack15 |
By ranking, do you mean reviews, or number of stars? |
| 12:48 |
MTDiscord |
<wsor4035> https://content.luanti.org/help/top_packages/ |
| 12:48 |
bgstack15 |
My first thought is to manipulate this cdb-proxy to just use an alias for each backend, e.g., cdb=https://content.luanti.org/, and then just prepending "cdb/" to each username. |
| 12:49 |
bgstack15 |
so it'd be like looking for "cdb/wuzzy/coolpackagename" |
| 12:49 |
bgstack15 |
instead of just "wuzzy/coolpackagename" |
| 12:49 |
MTDiscord |
<wsor4035> might work, but i have doubts |
| 12:50 |
bgstack15 |
oh yeah, it's all just BS right now. I've run out of creative things to do in the game, lol |
| 12:51 |
bgstack15 |
And wsor4035, since you asked, my use case is to easily provide my own mods that you have been struggling to get set up in the official ContentDB instance. I want easy access to official content and my own instance. |
| 12:54 |
bgstack15 |
ooh, and I've just had a thought. The client could even configure the cdbp (cdb-proxy) just with query parameters. I'm thinking, https://contentdbproxy.example.com/?cdb=https%20%23%23content.luanti.org&internal=https%20%23%23localhost%308080%23 |
| 12:54 |
bgstack15 |
And yes, I realize I didn't get the url encoding right on those. |
| 12:54 |
repetitivestrain |
another bug, coordinates provided to AreaStore queries are truncated into shorts with wraparound arithmetic |
| 12:55 |
MTDiscord |
<wsor4035> seems your straying into forking the client as well to support your proxy |
| 12:55 |
bgstack15 |
oh, does the client not support configuring the url of the contentdb it reaches out to? |
| 12:55 |
[MatrxMT] |
<Blockhead256> it does |
| 12:56 |
MTDiscord |
<wsor4035> you can change the url, but making it support your own api spec requires forking |
| 12:56 |
repetitivestrain |
so that AreaStore::get_areas_in_area (vector.new (-32768, -32768, -32768), vector.new (32768, 32768, 32768), ...) (initially a typo in my code) produces no results whatever because it is actually interpreted as a query between vector.new (-32768, -32768, -32768) and vector.new (32768, 32768, 32768) |
| 12:56 |
repetitivestrain |
vector.new (-32768, -32768, -32768), vector.new (-32768, -32768, -32768)* |
| 12:56 |
[MatrxMT] |
<Blockhead256> I think the debian people complained that nonfree stuff exists on the ContentDB so the default config of then-Minetest was technically serving nonfree software from a free package |
| 12:56 |
repetitivestrain |
don't you agree that this is not only confusing but also not documented anywhere |
| 12:57 |
bgstack15 |
Hm, I don't want to change the api spec. Obviously because then clients would have to be modified to support it. So this will take more thought. |
| 12:58 |
bgstack15 |
Blockhead256: yes, that is a very Debian thing to do. So that's part of the target use case, as well as like those 2 or 3 folks whose weed mods got removed due to the UK laws which govern CDB. |
| 12:59 |
bgstack15 |
I can stop theorizing here if it'll spam everybody here. I have appreciated the input so far though! |
| 12:59 |
MTDiscord |
<wsor4035> i forget what, but the debian wiki had nonsense about cdb for the longest time |
| 12:59 |
bgstack15 |
wsor4035: everybody knows Debian stuff lags forever. Arch wiki is where it's at, and I don't even use Arch. |
| 12:59 |
MTDiscord |
<wsor4035> also was extremely ironic as they had some mods as packages that where outdated af with license violations |
| 13:00 |
bgstack15 |
wsor4035: I realize you are a volunteer with limited time, et al, but when you get a chance, please review that conversation about my mod. I have some more info there. |
| 13:01 |
MTDiscord |
<wsor4035> i saw it last night, crafty - i was to tired to dot all the i's and cross all the t's last night. probably tonight |
| 13:03 |
bgstack15 |
Oh, yeah, so back to my <instancealias>/<author>/<name> thing, the proxy would merely manipulate all package identifiers by adding <instancealias>/, so the client would merely be told that "<instancealias>/<author>" is the author name. So yes, it would be exploiting that hopefully it doesn't strip out slashes from object names (which flask probably does, oh boy, there goes my plan). |
| 13:03 |
bgstack15 |
or maybe just <instancealias>.<author> |
| 13:03 |
bgstack15 |
I'll poke at it. |
| 13:07 |
[MatrxMT] |
<Blockhead256> if it strips slashes use another character :) Problem averted for now |
| 13:07 |
bgstack15 |
Well, I want to make sure usernames cannot have whatever character I pick. |
| 13:27 |
|
sknebel joined #luanti |
| 14:04 |
|
gregon joined #luanti |
| 14:05 |
|
gregon left #luanti |
| 14:08 |
|
nekobit joined #luanti |
| 14:17 |
|
simon816 joined #luanti |
| 14:18 |
rubenwardy |
bgstack15: yeah I've been thinking about a `cdblite` which is a brand new backend implementation that allows (1) getting content from upstream instances (2) providing comment from local json files (could be version controlled) |
| 14:18 |
rubenwardy |
the docs for this is https://github.com/luanti-org/contentdb/blob/master/docs/luanti_client.md |
| 14:19 |
bgstack15 |
thanks, rubenwardy! I found that eventually. An old link to a minetest_client.md is lurking somewhere in ContentDB [static pages, presumably] |
| 14:20 |
bgstack15 |
I fully expect my code is going to be garbage compared to yours, so maybe we should have a meeting where you explain what architecture you'd like to see and maybe I can get started working on something that meets your expectations. |
| 14:22 |
rubenwardy |
It can be very simple. My recommendation is to make it as easy to deploy as possible |
| 14:22 |
bgstack15 |
I'm stuck atm just trying to get a dev instance up without needing docker and postgres. *sigh* Maybe I should just use those. |
| 14:22 |
rubenwardy |
so a single application, no docker or db, cache things to filesystem/memory |
| 14:22 |
bgstack15 |
I don't want my proxy to need a db, yeah |
| 14:22 |
rubenwardy |
I was thinking making it with go |
| 14:22 |
rubenwardy |
(I don't have time for this) |
| 14:22 |
bgstack15 |
Oof, that would be a stretch for me. I'm more of a py/flask guy. |
| 14:22 |
rubenwardy |
(I don't have time to make cdblite, was hoping someone would do it) |
| 14:22 |
rubenwardy |
haha same |
| 14:22 |
rubenwardy |
would have been an excuse to learn go |
| 14:23 |
bgstack15 |
I also know nothing about caching stuff to disk/memory. |
| 14:23 |
bgstack15 |
Also I've just learned that my idea of using query params in the client's contentdb_url will not work (at first glance) because the client then adds query params for further requests. |
| 14:24 |
bgstack15 |
so it's more of a contentdb_hostname and not a _url |
| 14:24 |
rubenwardy |
you could do it as part of the URL path itself |
| 14:24 |
rubenwardy |
proxy.com/upstream_domain+upstream_domin/ |
| 14:25 |
rubenwardy |
plus is a space but looks cool encoded |
| 14:25 |
luatic |
rubenwardy: agreed on go. unfortunately at the moment i don't really have time either :p |
| 14:26 |
rubenwardy |
go has the benefit of being able to compile to a single distributable binary quite easily afaik |
| 14:26 |
luatic |
yep |
| 14:27 |
luatic |
should also consider something similar for the serverlist so schools can easily set up a proxy that whitelists a selection of servers |
| 14:29 |
rubenwardy |
My recommendation for schools is to disable the serverlist and add any servers to user favourites |
| 14:29 |
bgstack15 |
I would like to see the proxy/cdblite be configurable so that it either lets the client send its desired backends, or it can ignore those and use the preconfigured list. |
| 14:30 |
rubenwardy |
actually maybe we should read a share file first |
| 14:30 |
bgstack15 |
We can't solve the whole world's Internet trust problems, but we can make our tool configurable! |
| 14:30 |
rubenwardy |
to make it easier to distribute saved servers |
| 14:30 |
bgstack15 |
Ah, I liked being able to set up my own serverlist! |
| 14:43 |
|
aptgetefe joined #luanti |
| 14:43 |
|
aptgetefe joined #luanti |
| 15:06 |
|
citrons joined #luanti |
| 15:14 |
|
silverwolf73828 joined #luanti |
| 15:18 |
|
kimapr joined #luanti |
| 15:35 |
|
Kimapr_ joined #luanti |
| 15:38 |
|
Kimapr_ joined #luanti |
| 15:41 |
|
kimapr joined #luanti |
| 15:47 |
|
simon816 joined #luanti |
| 16:01 |
|
mrkubax10 joined #luanti |
| 16:06 |
|
Kimapr_ joined #luanti |
| 16:26 |
|
Kimapr_ joined #luanti |
| 16:34 |
|
Kimapr_ joined #luanti |
| 16:41 |
|
Kimapr_ joined #luanti |
| 16:41 |
[MatrxMT] |
<birdlover32767> you could have instructions on a certain file to disable features by removing code |
| 16:41 |
|
___nick___ joined #luanti |
| 16:47 |
|
kimapr joined #luanti |
| 16:54 |
|
kimapr joined #luanti |
| 17:02 |
|
Kimapr_ joined #luanti |
| 17:19 |
|
Talkless joined #luanti |
| 18:09 |
|
dabbill joined #luanti |
| 18:14 |
|
gregon joined #luanti |
| 18:15 |
|
gregon left #luanti |
| 18:39 |
|
SFENCE joined #luanti |
| 18:47 |
|
jaca122 joined #luanti |
| 18:52 |
|
SFENCE joined #luanti |
| 18:56 |
|
imi joined #luanti |
| 19:07 |
|
turtleman joined #luanti |
| 19:17 |
|
alias joined #luanti |
| 19:19 |
bgstack15 |
I've got a cruddy flask app that loops over the provided requested_backends, and browsing the content works, including caching the images on the client side! Unfortunately, I am stuck when I have it try to download. |
| 19:19 |
bgstack15 |
I have the exact same content length as the backend contentdb, and exact same payload and same Content-Type. I cannot get --trace to get any more verbose on the client to show me what's wrong. |
| 19:31 |
|
mrkubax10 joined #luanti |
| 19:36 |
|
aptgetefe joined #luanti |
| 19:44 |
|
kimapr joined #luanti |
| 19:51 |
|
Kimapr_ joined #luanti |
| 19:57 |
|
kimapr joined #luanti |
| 20:03 |
|
coco joined #luanti |
| 20:07 |
|
kimapr joined #luanti |
| 20:07 |
|
coco left #luanti |
| 20:11 |
sfan5 |
try wireshark |
| 20:14 |
|
SFENCE joined #luanti |
| 20:27 |
|
erle joined #luanti |
| 20:28 |
|
fbievan joined #luanti |
| 20:39 |
|
Kimapr_ joined #luanti |
| 20:40 |
erle |
Kimapr_ i have recently suggested to someone who likes riddles to play skyhell. let's see if you get more complai^W compliments! |
| 20:41 |
Kimapr |
!! |
| 20:41 |
Kimapr |
i should probably get into nodecore sometime, i haven't really played it much at all lately |
| 20:43 |
bgstack15 |
Haha, sfan5, I finally figured it out. The contents were exactly correct. However, because of the author name manipulation I'm doing, where the package it searched was "bgstack15~http___vm4_/mcl_colored_chests" but it properly got back "bgstack15/mcl_colored_chests" the client was not finding the package in the response that it looked for in the request. |
| 20:48 |
erle |
Kimapr i haven't figured out much nodecore actually. it's one of the best games though IMO. |
| 20:48 |
erle |
Kimapr did you ever play hyperrogue? it's very nodecorian |
| 20:48 |
Kimapr |
i did play hyperrogue |
| 20:49 |
erle |
cataclysm dda is also kinda neat |
| 20:50 |
erle |
because it has this energy of “if you assume video game tropes, you die” |
| 20:50 |
erle |
i wish there was a minetest/luanti game that did that |
| 20:50 |
erle |
exile is not that |
| 20:52 |
|
Kimapr_ joined #luanti |
| 20:53 |
Kimapr_ |
hmm, did my message fail to send? |
| 20:53 |
Kimapr_ |
erle: probably won't get in much detail here as it's offtopic but i am recently obsessed with bytebeat music and made a ffmpeg demuxer for the javascript variety, i'm pretty sure i saw your name when browsing some bytebeat things |
| 20:54 |
erle |
Kimapr yes i wrote a library for it and held a talk. long time ago! |
| 20:54 |
erle |
i once had a coworker tell me that he did not recognize me from the video, but i arguably do look very different |
| 20:59 |
erle |
Kimapr my personal opinion is that the code that you run to get a single sample should not be turing-complete so you can not have a halting problem when wanting to have a sample, but i think i am pretty much alone with that opionion, except for the author of the bytebeat ios app “glitch machine” (which libglitch basically emulates based on entirely legitimately acquired source code files) |
| 20:59 |
erle |
legitimate salvage! |
| 21:00 |
|
SFENCE joined #luanti |
| 21:01 |
erle |
does anyone else have an MNT reform 2 and has performance tips other than “disable shaders” and “enable mipmaps”? i know the engine is wasteful, but i kinda think the game runs way worse than it should on the hardware. |
| 21:01 |
erle |
like a laptop from 2021 should not give me worse performance than one from 2006 (but somehow it does) |
| 21:02 |
erle |
i do use xwayland though, does that affect 3d performance? |
| 21:04 |
bgstack15 |
I got the MVP for ContentDBProxy completed. I'm going to clean it up and dump it up on codeberg within the next few days. Technically I've only tried it when listing 2 instances, and not more. |
| 21:05 |
erle |
bgstack15 oh, what does it do? |
| 21:06 |
bgstack15 |
erle, it lets you set a contentdb_url that is crafted with multiple backend hostnames, so that you can point to more than one ContentDB in the client. It doesn't require any changes to the client other than the config entry. |
| 21:06 |
pgimeno_ |
repetitivestrain: it would be a shame if your comment is lost and forgotten, maybe post it in #luanti-dev or in a GH issue |
| 21:06 |
bgstack15 |
For the very few people that would want multiple ContentDB instances, such as a school one AND the public one. |
| 21:06 |
sfan5 |
generally good advice is to use a recent luanti version but that's not compatible with "disable shaders" advice |
| 21:07 |
sfan5 |
but also you can buy underpowered ARM hardware from 2021 and it will not perform better than average x86 hardware from 2006 |
| 21:07 |
bgstack15 |
But in general it would facilitate having an "unofficial" central ContentDB. |
| 21:07 |
erle |
i see, it is again, my fault and not the fault of people removing functions i use! |
| 21:07 |
bgstack15 |
(BYOD) |
| 21:07 |
erle |
i was not asking for snark though, but advice |
| 21:08 |
erle |
bgstack15 this vaguely reminds me of the fact that soemone (cora?) once showed me a cdb backend that was entirely static (which does need a modified client), so it could be hosted on random static hosts and mixed-and-matched |
| 21:08 |
erle |
bgstack15 unfortunately i remember very little, or else i would give you a link |
| 21:08 |
sfan5 |
erle: I do not have that hardware so I can't give specific advice beyond that |
| 21:09 |
bgstack15 |
That does sound interesting, erle, but in a different direction that the inspiration for my current project. |
| 21:10 |
erle |
bgstack15 obviously, yours does not need a modified client. but on anarchy servers, modified clients are more common. fleckenstein even made cheatDB, which was a contentdb for client side mods. |
| 21:11 |
erle |
it kinda failed to get traction though, because 1. it turns out client side mods are intimately coupled to the APIs that alternative clients expose 2. you probably want all the client side mods you can get, it's only a couple dozen per unofficial client anyway |
| 21:11 |
bgstack15 |
I'm much more a server guy than a desktop client compiler guy, so I tend to solve all my problems with more docker instances or python flask servers running somewhere. |
| 21:11 |
erle |
i can relate to that, even though i used to use bottle.py instead of flask |
| 21:11 |
erle |
it's great if you can change a backend and the frontend stays the same |
| 21:11 |
bgstack15 |
I use the Luanti flatpak like a noob |
| 21:12 |
erle |
spotted the fedora user! |
| 21:12 |
erle |
:P |
| 21:12 |
bgstack15 |
Hm, I used to be but now I run Devuan. |
| 21:12 |
erle |
damn |
| 21:12 |
erle |
i think the most impressive thing i have been shown on the server side was himbeerservers server multiplexer |
| 21:14 |
erle |
bgstack15 out of curiosity: what is the use case for having multiple cdbs? private mods that are only used in a school/university and need to be published according to other rules? |
| 21:17 |
erle |
like yes i read what you wrote before, but is it more “just in case” or do you have a current use case? |
| 21:19 |
|
Kimapr_ joined #luanti |
| 21:35 |
|
SFENCE joined #luanti |
| 21:41 |
erle |
when i am standing at an angle to a wall and it stops being rendered at the screen edges, what is the setting that adjusts that? |
| 21:41 |
sfan5 |
uhh |
| 21:41 |
sfan5 |
near plane? |
| 21:42 |
erle |
thx |
| 21:43 |
erle |
the effect i have right now is that when looking at the wall directly it renders, when i look 90 degrees to the side it renders, but when i look at it between 30 to 75 degrees turned a part of it is not rendered (and i can see behind) |
| 21:43 |
erle |
i will adjust the setting and report back |
| 21:43 |
sfan5 |
if your device has a broken GL driver performance would be the least of my worries |
| 21:45 |
erle |
i have so far only noticed two things that are definitely broken with the driver: 1. the green gear in glxgears has no hole 2. in neverball all ground textures are black. but minetest/luanti works fine. |
| 21:46 |
erle |
what's the technical name of the near plane setting? i can't seem to find it |
| 21:46 |
|
kimapr joined #luanti |
| 21:47 |
erle |
sfan5 the reason i was asking is also because i have had this in the past with e.g. upright sprites and it was an engine issue, not a driver one |
| 21:48 |
sfan5 |
the near plane setting no longer exists in the engine |
| 21:48 |
erle |
and like with the upright sprites it only happens when i am *right* in front of the wall |
| 21:48 |
erle |
so i was guessing it's a similar bug |
| 21:53 |
erle |
sfan5 do you remember why it was removed? i mean it doesn't get cheatier than “i can have xray if i turn my head” but i don't get why e.g. it was still left enabled on android. |
| 21:55 |
erle |
i guess i have to live with unintentional xray |
| 21:57 |
erle |
ttyl |
| 22:04 |
|
Kimapr joined #luanti |
| 22:28 |
|
ineva joined #luanti |
| 22:33 |
|
panwolfram joined #luanti |
| 23:05 |
|
Eragon joined #luanti |
| 23:16 |
|
chilledfrogs joined #luanti |
