Time |
Nick |
Message |
00:00 |
paramat |
i would love that |
00:03 |
paramat |
it would fit with c55's roadmap too about universality |
00:03 |
rubenwardy |
yeah, exactly |
00:03 |
|
Void7 joined #minetest-dev |
00:03 |
rubenwardy |
ages ago I was talking to PilzAdam in PM about a gravity system |
00:03 |
rubenwardy |
Mods would register point masses |
00:04 |
rubenwardy |
and there could be linear gravity points |
00:04 |
rubenwardy |
for games like most which have a normal MT world |
00:05 |
rubenwardy |
gravity would be 9.81 at y=0, and decrease when going either direction |
00:05 |
rubenwardy |
although it would decrease more going up than going down |
00:05 |
rubenwardy |
as in, there would be less gravity at 30000 than at -30000 |
00:05 |
rubenwardy |
ideas are cheap though |
00:05 |
agrecascino |
how would i add a network opcode? |
00:06 |
paramat |
are you thinking gravity has 6 directions only? i think that's better |
00:06 |
rubenwardy |
agrecascino: https://github.com/minetest/minetest/blob/master/src/network/networkprotocol.h#L607 |
00:06 |
agrecascino |
thanks |
00:07 |
rubenwardy |
I think |
00:07 |
rubenwardy |
paramat, they're be two types of gravity. Linear gravity is for masses whose centres are off map, for example the main world. For things like moons there could be point gravity, so at the center |
00:08 |
rubenwardy |
problem with 6-dir gravity is the transition when at the edges |
00:08 |
rubenwardy |
by linear gravity I mean the direction is always down, because the center of the earth is "infinitely" down wards |
00:08 |
agrecascino |
i don't really understand 3d collision math, so i can't help you on this |
00:08 |
rubenwardy |
as compared to point gravity |
00:09 |
rubenwardy |
I understand the gravity equations |
00:09 |
rubenwardy |
I wrote a HTML5 sim: http://physics.rubenwardy.com/sims/gravity/gravity.html |
00:09 |
rubenwardy |
but it won't be as easy as that in MT due to how physics doesn't use forces |
00:11 |
paramat |
there's a voxel game called stellar overload which has cubic planets and 6-fold gravity |
00:12 |
paramat |
gravity flips at planet edges |
00:12 |
rubenwardy |
for cube gravity that would make sense |
00:12 |
rubenwardy |
I actually thing that makes the most sense in a voxel world |
00:13 |
rubenwardy |
although, what happens if the player erodes the edges? |
00:13 |
T4im |
wouldn't that still be point gravity, just made discrete? |
00:13 |
rubenwardy |
maybe make the cube big enough |
00:13 |
paramat |
gravity direction would be determined by diagonal planes emerging from centre |
00:14 |
|
betterthanyou710 joined #minetest-dev |
00:15 |
rubenwardy |
Another topic: what would be the most feasible way to add syntax highlighting to a formspec? |
00:15 |
rubenwardy |
Thinking about making an educational subgame to teach people how to program |
00:15 |
rubenwardy |
would also be useful for Lua controllers from mesecons |
00:16 |
paramat |
https://www.youtube.com/user/PlanetsCube/videos |
00:17 |
TheReaperKing |
rubenwardy: I'd be super interested |
00:17 |
agrecascino |
is there a command to force a client to unload an object? |
00:17 |
agrecascino |
opcode* |
00:17 |
rubenwardy |
TheReaperKing, for an educational subgame? |
00:17 |
TheReaperKing |
Some of my students and I have started to try to use them but haven't gotten them working |
00:18 |
TheReaperKing |
I would really love for them to be able to program with Lua in game |
00:18 |
TheReaperKing |
We haven't tried much, just to get that time sensor tutorial working |
00:18 |
TheReaperKing |
I did get it to print a message to the console correctly |
00:19 |
TheReaperKing |
We are still in the design unit so I haven't put that much time into it, but they'll definitely be learning lua one way or another |
00:19 |
TheReaperKing |
and me too hehe |
00:20 |
TheReaperKing |
I actually found this program that uses scratch like blocks and then it converts it to lua |
00:20 |
TheReaperKing |
so we can go scratch, blocky, to full on lua |
00:20 |
TheReaperKing |
or maybe even just start lua |
00:21 |
TheReaperKing |
this is the one I tried but couldn't get to work, could never get the LCD to display |
00:21 |
TheReaperKing |
http://mesecons.net/digilines.html |
00:26 |
TheReaperKing |
I'll be working on the other tutorials for sure though. I wonder if you could even save script files to file in game |
00:36 |
Hijiri |
how does the server keep track of what keys players are pressing? |
00:36 |
Hijiri |
for example, in get_control |
00:37 |
Hijiri |
I was thinking of how to implement "extra keybindings", where the player could bind "real keys" to keys like "extra 1", "extra 2", etc. The basic idea was described by Wuzzy on the forums |
00:38 |
Hijiri |
Then on top of that you could write a mod that would handle allocating the numerical keys to mods that request it |
00:38 |
Hijiri |
(or it could be part of the core lua) |
00:42 |
|
raingloom joined #minetest-dev |
00:44 |
Hijiri |
well keyPressed has room for 23 more keys |
00:44 |
Hijiri |
are there any plans to use them for anything? |
00:57 |
Hijiri |
actually maybe it would be better to have a separate packet type for extra key press/release events? |
00:58 |
Hijiri |
or update or whatever, in case you don't want the loss of a release event to break stuff |
00:59 |
Hijiri |
alternatively #3777 would be something compatible with devices that don't use the keyboard |
00:59 |
ShadowBot |
https://github.com/minetest/minetest/issues/3777 -- API for adding arbitrary elements to the hotbar |
00:59 |
Hijiri |
It would be harder to implement though, but you could use item definitions for the extra elements |
01:09 |
|
Miner_48er joined #minetest-dev |
01:26 |
|
STHGOM joined #minetest-dev |
01:26 |
|
STHGOM joined #minetest-dev |
01:44 |
Hijiri |
also it wouldn't provide everything that controls do |
01:46 |
|
paramat joined #minetest-dev |
02:00 |
|
nnnn20430 joined #minetest-dev |
02:19 |
|
betterthanyou711 joined #minetest-dev |
02:21 |
|
ssieb joined #minetest-dev |
02:44 |
|
ssieb joined #minetest-dev |
02:58 |
|
Void7 joined #minetest-dev |
03:04 |
agrecascino |
paramat, do you know how minetest's networking works? |
03:07 |
paramat |
no, sorry |
03:21 |
|
Hasan2 joined #minetest-dev |
03:24 |
|
froike- joined #minetest-dev |
03:25 |
|
Twirlimp_ joined #minetest-dev |
03:25 |
|
thatgraemeguy_ joined #minetest-dev |
03:25 |
|
thatgraemeguy_ joined #minetest-dev |
03:26 |
|
friti joined #minetest-dev |
03:27 |
|
exoplanet joined #minetest-dev |
03:27 |
|
exoplanet joined #minetest-dev |
03:27 |
|
exoplanet joined #minetest-dev |
03:27 |
|
dmurph joined #minetest-dev |
03:29 |
|
VargaD joined #minetest-dev |
03:29 |
|
misprint joined #minetest-dev |
03:47 |
agrecascino |
does anyone understand the networking in minetest well? |
03:50 |
|
\o` joined #minetest-dev |
03:51 |
|
endev15 joined #minetest-dev |
03:56 |
|
octacian joined #minetest-dev |
03:58 |
paramat |
some devs do yes |
04:02 |
|
octacian joined #minetest-dev |
04:20 |
agrecascino |
alright |
04:20 |
agrecascino |
if anybody knows |
04:20 |
agrecascino |
how does damage and or object unloading work on serverside |
04:20 |
agrecascino |
thanks |
04:33 |
|
Miner_48er joined #minetest-dev |
05:13 |
|
paramat left #minetest-dev |
06:06 |
|
red-001 joined #minetest-dev |
06:06 |
|
Hunterz joined #minetest-dev |
06:56 |
|
red-001 joined #minetest-dev |
06:57 |
red-001 |
could something be done about #3900? |
06:57 |
ShadowBot |
https://github.com/minetest/minetest/issues/3900 -- Add a button for disabling all mods to world config. by red-001 |
07:17 |
|
damiel_ joined #minetest-dev |
07:49 |
|
Gundul joined #minetest-dev |
07:57 |
|
red-001 joined #minetest-dev |
08:02 |
|
davisonio joined #minetest-dev |
08:18 |
|
sofar joined #minetest-dev |
08:20 |
|
Krock joined #minetest-dev |
09:13 |
|
damiel_ joined #minetest-dev |
09:55 |
|
T4im joined #minetest-dev |
10:08 |
|
Darcidride joined #minetest-dev |
10:19 |
|
Fixer joined #minetest-dev |
10:25 |
|
ensonic joined #minetest-dev |
10:26 |
ensonic |
hi, can a texturepack override textures from a mod? |
10:27 |
Krock |
yes, that's its purpose |
10:27 |
Krock |
it can also replace model textures |
10:28 |
* ensonic |
is hacking on mcresconvert and mcimport |
10:28 |
ensonic |
the mine-trek voyager already looks quite good |
10:29 |
Krock |
make it so, Chackotey. |
10:34 |
ensonic |
:) |
10:48 |
|
nrzkt joined #minetest-dev |
11:29 |
|
betterthanyou710 joined #minetest-dev |
12:41 |
|
AnotherBrick joined #minetest-dev |
12:43 |
|
davisonio joined #minetest-dev |
12:44 |
|
rubenwardy joined #minetest-dev |
12:55 |
|
red-001 joined #minetest-dev |
13:26 |
|
raingloom joined #minetest-dev |
13:26 |
|
Robby joined #minetest-dev |
13:35 |
|
STHGOM joined #minetest-dev |
13:35 |
|
STHGOM joined #minetest-dev |
13:42 |
Fixer |
!seem MillersMan |
13:42 |
Fixer |
!seen MillersMan |
13:42 |
ShadowBot |
Fixer: I saw MillersMan in #minetest-dev 8 weeks, 0 days, 14 hours, 25 minutes, and 32 seconds ago saying "Maybe the wrong time to ask this but: Is there any interest in having a generic base-implementation for mapgens that work similar to my Canyon mapgen?" |
13:43 |
Fixer |
crap, now water reflow is in limbo too |
13:45 |
red-001 |
water reflow? |
13:45 |
red-001 |
related to the ignore bugs? |
13:46 |
red-001 |
is that the pr that reflowed the water on block load? |
13:46 |
red-001 |
!seen red-001 |
13:46 |
ShadowBot |
red-001: I saw red-001 in #minetest-dev 0 seconds ago saying "!seen red-001" |
13:47 |
red-001 |
!seen davisonio |
13:47 |
ShadowBot |
red-001: I saw davisonio in #minetest-dev 7 weeks, 3 days, 18 hours, 35 minutes, and 34 seconds ago saying "seems like the docs are at https://lgtm.co/docs/install/ instead" |
13:47 |
red-001 |
!seen book` |
13:47 |
ShadowBot |
red-001: I haven't seen book` in #minetest-dev. |
13:47 |
red-001 |
!seen book |
13:47 |
ShadowBot |
red-001: I haven't seen book in #minetest-dev. |
14:05 |
|
Wuzzy joined #minetest-dev |
14:06 |
Wuzzy |
I noticed something odd in Minetest Game 0.4.14: The xpane nodes have buildable_to set to true |
14:06 |
Wuzzy |
but in-game, they don't behave like buildable_to normally do: building does not replace them, and falling nodes do not go through and destroy them |
14:07 |
Wuzzy |
anyone knows why buildable_to was set to true for these nodes? |
14:10 |
Wuzzy |
actually, the whole pane_def_fields in the xpanes mod seems very odd: drawtype = "airlike", walkable = false, etc. |
14:23 |
|
blaze joined #minetest-dev |
14:31 |
|
Void7 joined #minetest-dev |
14:36 |
|
KaadmY joined #minetest-dev |
14:41 |
|
hmmmm joined #minetest-dev |
14:49 |
|
pperron joined #minetest-dev |
14:49 |
rubenwardy |
Does anyone know about service broadcasting over LAN? I was thinking about whether it could be possible to auto detect servers on your LAN network and show them in the server list |
14:49 |
rubenwardy |
looks like this might be what I'm looking for: https://en.wikipedia.org/wiki/Simple_Service_Discovery_Protocol |
14:50 |
rubenwardy |
usecase: to make it easier for technonoobs to setup a local server, they won't need to open a terminal an find their ip then |
14:50 |
rubenwardy |
*and |
14:52 |
|
raingloom joined #minetest-dev |
14:53 |
celeron55 |
it has been talked before and is just a matter of making the client send some broadcast packets |
14:53 |
ensonic |
rubenwardy, the easiest 'solution' is to put a static xml file into /etc/avahi/services/ (on linux) |
14:54 |
celeron55 |
and making the server understand them and respond to them |
14:55 |
ensonic |
rubenwardy, but we'd need to define a service type so that clients would discover server annoucements |
14:55 |
ensonic |
this would be using mdns |
14:55 |
rubenwardy |
celeron55: would you be able to give me some names to research? |
14:55 |
ensonic |
celeron55, why invent some new discovery mechanism? |
14:57 |
|
twoelk joined #minetest-dev |
14:57 |
ensonic |
here are some sample xml files for static announcements: https://wiki.archlinux.org/index.php/avahi#NFS, doing this dynamically when the server starts is much nicer of course |
14:57 |
celeron55 |
ensonic: windows? |
14:57 |
celeron55 |
ensonic: android? |
14:57 |
celeron55 |
ensonic: something else than linux? |
14:57 |
ensonic |
celeron55, zeroconf/bonjour works on all those platforms |
14:58 |
ensonic |
celeron55, the static conf was only meant as an example |
14:58 |
rubenwardy |
a solution should be as automatic as possible - the server owner is likely to just be hosting using the server tab |
14:58 |
celeron55 |
well i don't have time to research this but if it's something that every user can use, then whatever |
14:59 |
celeron55 |
it should work out of the box on every platform |
14:59 |
celeron55 |
udp does, that i do know |
15:00 |
ensonic |
rubenwardy, the server would announce itself via broadcast when the start and upon scan requests from clients, client would when starting send a disover req and add minetest server from the local net to the server list |
15:01 |
rubenwardy |
ahh, there's a broadcast address on the network! |
15:01 |
rubenwardy |
that makes things a lot easier |
15:01 |
ensonic |
celeron55, mdns uses udp, it is quite close to dns (udp/53) but uses (udp/5353) - it is also not very complicated, we could have a built-in fallback, I wrote mdns stacks in js in not too many lines of code :) |
15:02 |
rubenwardy |
it would be as simple as the client sending a packet (or a few incase the first fails due to udp) and the server listening for broadcast packets, and sending a "hello! I'm here!" packet to the client when it hears one |
15:04 |
rubenwardy |
now I just need to try and setup a virtual LAN network using VMs so I can test... |
15:09 |
twoelk |
would the client only search on start up? maybe each time the serverlist is called or have a search now button? |
15:16 |
|
agrecascino joined #minetest-dev |
15:25 |
|
Void7 joined #minetest-dev |
15:30 |
|
betterthanyou710 joined #minetest-dev |
15:37 |
Calinou |
there should be a refresh button in server list anyway |
15:46 |
agrecascino |
anyone here understand minetest's networking? |
15:47 |
Wuzzy |
Is it a bug that falling nodes can go through walkable nodes if the walkable node has buildable_to=true? |
15:54 |
|
Void7 joined #minetest-dev |
16:02 |
red-001 |
#4472 |
16:02 |
ShadowBot |
https://github.com/minetest/minetest/issues/4472 -- Make serverlist searchable. by red-001 |
16:06 |
|
betterthanyou710 joined #minetest-dev |
16:12 |
twoelk |
uhm load http://servers.minetest.net/list and use your browsers search function |
16:12 |
Krock |
not very handy on smartphones |
16:21 |
red-001 |
twoelk might as well remove the serverlist |
16:21 |
red-001 |
after all you can just copy the port and address |
16:22 |
twoelk |
in contrast to their german colloquial name I don't think mobile phones are really handy anyways |
16:24 |
red-001 |
I thought shadowbot generated different links for prs and issues? |
16:24 |
red-001 |
when did this change? |
16:26 |
red-001 |
e.g it used github/minetest/minetest/pull/ for pr's and github.com/minetest/minetest/issue for issues |
16:26 |
rubenwardy |
I don't think it matters |
16:27 |
rubenwardy |
How do I include changes from upstream/master into my origin/patch-1? Is it git pull --rebase upstream master |
16:27 |
rubenwardy |
I can never remember |
16:28 |
celeron55 |
if you want to "move" the patch commits to come after master commits, that would do it |
16:28 |
rubenwardy |
yeah, that's what I want |
16:28 |
celeron55 |
umm... maybe |
16:28 |
rubenwardy |
I'll back up just in case |
16:28 |
celeron55 |
i actually never use pull --rebase |
16:29 |
celeron55 |
i just looked up the man page and yes, that's what it does |
16:29 |
|
blaze joined #minetest-dev |
16:29 |
celeron55 |
...and i think i should use pull --rebase more often |
16:29 |
red-001 |
rubenwardy it can be useful to know if something is a pr or issue without opening it |
16:30 |
|
nrzkt joined #minetest-dev |
16:33 |
red-001 |
is https://github.com/orgs/minetest/people the list of core devs? |
16:35 |
|
alkotob joined #minetest-dev |
16:35 |
Krock |
yes |
16:36 |
Krock |
make sure it says "Unfollow" on each button on the right side |
16:36 |
red-001 |
:) |
16:36 |
|
blaze left #minetest-dev |
16:39 |
Krock |
red-001, the typo monster doens't like you: "[...] that are simultaneously send pre client." |
16:39 |
Krock |
pre |
16:40 |
rubenwardy |
red-001, apart from Calin_u, he's a web dev |
16:41 |
Krock |
also, "to send" is in this case with "to be" ("are") -> "sent" |
16:41 |
rubenwardy |
I think that's it |
16:41 |
Krock |
I'm not a native english speaker so shoot at me when I'm wrong |
16:45 |
rubenwardy |
*shoot me if I'm wrong |
16:46 |
rubenwardy |
just kidding, yours was fine |
16:50 |
* celeron55 |
measures darkness of humour, drops the result under the bed and can't find it anymore |
16:52 |
Krock |
^^ |
16:53 |
Fixer |
639 players |
18:08 |
|
Grandolf joined #minetest-dev |
18:09 |
|
halt_ joined #minetest-dev |
18:13 |
|
Void7 joined #minetest-dev |
18:32 |
|
paramat joined #minetest-dev |
18:42 |
|
agrecascino joined #minetest-dev |
19:00 |
|
damiel_ joined #minetest-dev |
19:10 |
APNG |
can we get player-defined player models? |
19:10 |
APNG |
and unlockable flight/etc capabilities, through dungeon items? |
19:11 |
|
agrecascino joined #minetest-dev |
19:11 |
APNG |
mostly because that'd be furry-friendly |
19:11 |
APNG |
and furries are one of the most welcoming communities, so we wouldn't have so many assholes in our community |
19:18 |
sofar |
chat about furries is offtopic |
19:19 |
APNG |
it's about furries sure, it's also about the minetest community, the users, and even the developers >.> |
19:19 |
sofar |
unless it's about game or game engine development, it's off topic |
19:19 |
APNG |
so #minetest-dev doesn't care about minetest? |
19:19 |
APNG |
good, fucking useless >.> |
19:19 |
sofar |
you love to twist words, don't you |
19:20 |
sofar |
nobody can have an honest discussion with you since now I've apparently raped furries |
19:20 |
twoelk |
this is not about furrieys but about the easiness to change the player model |
19:21 |
sofar |
THEN WHY THE FUCK DID YOU MENTION FURRIES? |
19:21 |
sofar |
sorry, that was uncalled for |
19:21 |
sofar |
please, stay technical in chat in here? |
19:21 |
APNG |
so #minetest-dev doesn't care about the minetest community? |
19:21 |
sofar |
you can't have player defined models since the server needs to know all the models at server start |
19:22 |
APNG |
which, fyi, is more important than the game code |
19:22 |
sofar |
you can have player defined models if that means that the player chooses a model from a list of known models |
19:22 |
sofar |
you can have items that grant flight or abilities, just write a mod that does that |
19:22 |
APNG |
but how would it work? |
19:22 |
APNG |
you see |
19:23 |
sofar |
you'd write some lua code |
19:23 |
APNG |
granting flight or w/e only grants the same thing everyone has |
19:23 |
APNG |
it doesn't let you control flight semantics or w/e |
19:23 |
APNG |
so there's no good UX |
19:23 |
sofar |
you'd have to write a mod to make whatever UX you want |
19:23 |
sofar |
e.g. helicopter mod makes a specific UX in the form of a flying vehicle |
19:24 |
APNG |
mods can't control how high you can fly or how you do the flying without breaking some things |
19:24 |
sofar |
file a bug? |
19:24 |
APNG |
a mod can't make it so you have to hold space to go up, stop holding space to go down, etc |
19:25 |
APNG |
or tap space repeatedly to go up (like wings) |
19:25 |
APNG |
and limiting the fly height would leak into things |
19:25 |
sofar |
you could do "flapping" with an item in hand and repeatedly using it to "flap" in a mod |
19:26 |
APNG |
also, fly has to be manually activated in the config, instead of being something that can be toggled on/off with an item |
19:26 |
sofar |
all movement buttons are capturable in mods too |
19:26 |
sofar |
no, you can make people fly even if flying is off in the config |
19:26 |
sofar |
all player movement can be defined and overridden in mods |
19:27 |
APNG |
but doesn't that have to go through the network? |
19:27 |
APNG |
sounds wonky |
19:27 |
sofar |
? |
19:28 |
sofar |
everything has to go through the network at some point |
19:28 |
APNG |
yes, but I mean |
19:28 |
APNG |
movement can be simulated on the client |
19:28 |
APNG |
except mods can't simulate movement on the client |
19:28 |
APNG |
so instead you have to send the keys through the network, which might take 500 ms |
19:29 |
APNG |
then the mod has to tell the client what to do, which takes another 500 ms |
19:29 |
sofar |
minetest is a network game |
19:29 |
sofar |
that's not going to change |
19:29 |
APNG |
next thing you know it takes 1 second to start flying and another second to stop flying |
19:29 |
sofar |
minetest is a network game.... again |
19:29 |
sofar |
you can't get rid of that problem, it's part of the design |
19:29 |
APNG |
sofar, yes, but at least it could do some things on the client to make it less shitty |
19:29 |
APNG |
you can get rid of that problem |
19:30 |
APNG |
minecraft doesn't have that problem |
19:30 |
|
agrecascino_ joined #minetest-dev |
19:30 |
agrecascino_ |
hey APNG you like bikesheds |
19:30 |
sofar |
it has other problems |
19:30 |
agrecascino_ |
i do too ;) |
19:30 |
sofar |
back to your problem, you can do it in mods, problem seems solved? |
19:31 |
APNG |
eh |
19:31 |
APNG |
when do we get client-side mods :/ |
19:31 |
sofar |
when you write the code, or someone else does |
19:31 |
agrecascino_ |
i could make a logo for it |
19:32 |
APNG |
ok |
19:32 |
APNG |
in more pressing matters |
19:32 |
sofar |
yes, a logo, that would really help development of that code... |
19:32 |
APNG |
can we get end-to-end encrypted PMs? |
19:32 |
sofar |
sure, install signal messenger on your phone |
19:33 |
sofar |
oh, you mean in minetest? |
19:33 |
sofar |
not going to happen. |
19:33 |
APNG |
yes, in minetest >.> |
19:34 |
|
agrecascino joined #minetest-dev |
19:34 |
agrecascino |
sofar, is there already infrastructure for running lua on clients? |
19:34 |
agrecascino |
or is there none? |
19:34 |
Calinou |
agrecascino: there is literally none |
19:35 |
agrecascino |
oh god |
19:35 |
sofar |
client stays 100% in C++ atm. |
19:35 |
agrecascino |
how is mruby's performance compared to luajit? |
19:35 |
Calinou |
mruby? |
19:37 |
agrecascino |
Calinou, basically ruby for integration |
19:37 |
Fixer |
ruby has jit? |
19:38 |
sofar |
ruby wouldn't be a choice for client side mods anyway |
19:38 |
sofar |
it's going to be lua no matter what |
19:38 |
agrecascino |
sofar, yeah, since everything else is written in lua |
19:38 |
sofar |
even C++ mods won't make any sense |
19:38 |
agrecascino |
? |
19:38 |
agrecascino |
sofar, i disagree |
19:39 |
Calinou |
Ruby is slow compared to LuaJIT |
19:39 |
Calinou |
Ruby is not JIT, and no JIT implementations seem to exist for Ruby |
19:39 |
Calinou |
(unlike Python) |
19:39 |
agrecascino |
is pypy a jit? |
19:39 |
sofar |
explain how an x86_64 server is going to send a precompiled binary to an armv7 android client then |
19:39 |
Calinou |
agrecascino: yes |
19:39 |
Calinou |
sofar: cross-compilation :P |
19:39 |
Calinou |
but yeah, C++ mods are a bad idea |
19:39 |
agrecascino |
sofar, why not use something like cling |
19:40 |
Calinou |
even with NaCl you still need to compile for every architecture |
19:40 |
sofar |
you'd have to precompile possibly dozens of versions |
19:40 |
sofar |
it's not worth it |
19:40 |
agrecascino |
and compile on the client |
19:40 |
sofar |
lua solves the whole problem |
19:40 |
|
Fixer joined #minetest-dev |
19:40 |
Calinou |
(NaCl is platform-, but not architecture-independent) |
19:41 |
sofar |
plus c++ mods from a server could hack clients and are inherently insecure |
19:41 |
agrecascino |
sofar, what about something like a pledge call? |
19:41 |
sofar |
a what? |
19:41 |
Calinou |
you mean a crowdfunding campaign? |
19:41 |
agrecascino |
http://man.openbsd.org/cgi-bin/man.cgi/OpenBSD-current/man2/pledge.2?query=pledge |
19:42 |
sofar |
agrecascino: how is a BSD specific thing going to help linux or windows clients? |
19:42 |
sofar |
again, native mods are a huge security problem |
19:42 |
agrecascino |
sofar, i was thinking the idea, not the exact implementation |
19:43 |
T4im |
Calinou: JRuby is JIT |
19:43 |
Calinou |
T4im: in my experience it's not much faster than official Ruby |
19:43 |
Calinou |
can be slower in many cases |
19:43 |
Calinou |
unlike PyPy which is almost always faster than CPython |
19:43 |
|
nrzkt joined #minetest-dev |
19:44 |
Fixer |
too bad there is no good anticheat in minetest, I seen hackers who can fly/fast/teleport/noclip... |
19:44 |
Calinou |
Fixer: no good anticheat exists for any game |
19:45 |
Fixer |
agree |
19:45 |
agrecascino |
sofar, do you understand networking in mintest? |
19:45 |
Calinou |
the only solution is moderation |
19:45 |
agrecascino |
BattlEye rootkit for minetest! |
19:46 |
Fixer |
Calinou: at least some kind of noclip/fly/fast detector will be useful (presented as a report) |
19:46 |
sofar |
agrecascino: it's not that complex |
19:46 |
Fixer |
there are already hacked minetest clients with cheats |
19:46 |
agrecascino |
sofar, but it's poorly documented |
19:46 |
sofar |
the code is the documentation |
19:47 |
APNG |
<Calinou> even with NaCl you still need to compile for every architecture |
19:47 |
APNG |
asm.js |
19:47 |
Calinou |
not gonna happen, we ain't adding V8 or such in Minetest… |
19:47 |
agrecascino |
asm.js is still incredibly slow |
19:47 |
APNG |
seamonkey |
19:47 |
APNG |
fuck v8 |
19:47 |
APNG |
v8 is shit |
19:47 |
APNG |
v8 eats half the RAM to run a hello world |
19:47 |
agrecascino |
APNG, then what else is there? |
19:47 |
agrecascino |
duktape? |
19:47 |
APNG |
agrecascino, seamonkey |
19:48 |
sofar |
duktape isn't bad |
19:48 |
Calinou |
the speed concern remains |
19:48 |
APNG |
or w/e that mozilla thing is called |
19:48 |
agrecascino |
what |
19:48 |
Calinou |
SpiderMonkey? |
19:48 |
Calinou |
0 A.D. uses it, they seem to regret it |
19:48 |
agrecascino |
spidermonkey is sloooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooow. |
19:48 |
sofar |
but, why would you ever do JS? |
19:48 |
APNG |
Calinou, it's faster than non-JIT Lua |
19:48 |
agrecascino |
sofar, because they obviously hate themselves |
19:48 |
APNG |
sofar, because emscripten |
19:48 |
APNG |
it's literally the only reason you'd ever use JS |
19:49 |
agrecascino |
why not add c++ mods then |
19:49 |
APNG |
the modern web doesn't use JS, it uses emscripten |
19:49 |
agrecascino |
and somehow sanitize the code |
19:49 |
Fixer |
damn, what are they smoking for this names... |
19:49 |
APNG |
agrecascino, again, emscripten |
19:49 |
sofar |
I'm pretty sure this is a non-discussion, if we'll ever have client side mods they'll be lua |
19:49 |
Calinou |
client-side modding, if done, should be done in an interpreted language |
19:49 |
Calinou |
I do not want blobs being downloaded from a server and run on my machine |
19:50 |
agrecascino |
and somehow sanitize the code |
19:50 |
agrecascino |
oops |
19:50 |
agrecascino |
meant |
19:50 |
APNG |
Calinou, thus JS |
19:50 |
agrecascino |
"why are we even arguing this" |
19:50 |
APNG |
allows C and C++ mods |
19:50 |
Calinou |
APNG: we support one interpreted language already |
19:50 |
APNG |
Calinou, then give me an asm.lua >.> |
19:50 |
Calinou |
if by "allow", you mean "run at 10% speed with tons of restrictions", maybe |
19:50 |
Calinou |
also C/C++ are not good modding languages |
19:50 |
Calinou |
they make things much harder than they should be |
19:50 |
agrecascino |
and lua doesn't |
19:51 |
agrecascino |
one word |
19:51 |
agrecascino |
nil |
19:51 |
Fixer |
whats the problem with lua? tenplus1 run server with up to 40 servers, and lag was manageble, should not block inclusion of carts/mob in lua %) |
19:52 |
Calinou |
lag in Minetest is coming from the engine (server and client) |
19:52 |
Calinou |
and also because people expect too much with mods sometimes |
19:52 |
Calinou |
like, running tons of code every step |
19:52 |
Fixer |
server lag* |
19:53 |
paramat |
sofar rubenwardy game#1270 ? |
19:53 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/1270 -- Default, stairs, doors: Vary wood flammable and choppy group values by paramat |
19:54 |
APNG |
lag happens because LuaJIT can't JIT through C/Lua boundary |
19:54 |
APNG |
minetest should be reimplemented through the FFI |
19:54 |
agrecascino |
oh god |
20:07 |
Fixer |
what annoying thing is memory limit |
20:07 |
Fixer |
try darkage mod with some bunch of other mods, lua mem consumption... 30 40 50 40 20 30 50 80 40 50 20 LUA OOM |
20:09 |
Fixer |
suddenly |
20:09 |
|
red-001 joined #minetest-dev |
20:11 |
|
est31 joined #minetest-dev |
20:12 |
est31 |
I am a friend of 6-fold gravity |
20:13 |
est31 |
transition is very easy: calculate a vector of gravity |
20:13 |
est31 |
then take the closest of the 6 directions as "down" |
20:13 |
est31 |
and if the norm of the vector is smaller than some value, allow free movement or sth |
20:15 |
red-001 |
!seen loggingbot_ |
20:15 |
ShadowBot |
red-001: I haven't seen loggingbot_ in #minetest-dev. |
20:20 |
|
eeew` joined #minetest-dev |
20:24 |
APNG |
Fixer, swap luajit with lua 5.1 |
20:24 |
Fixer |
APNG: lua 5.1 is much slower |
20:24 |
Fixer |
APNG: and those lua oom errors can be fixed by writing mods properly |
20:24 |
APNG |
Fixer, but it can eat all your RAM |
20:25 |
Calinou |
Minetest supports plain Lua, but it's rarely used due to low performance :P |
20:26 |
APNG |
RAM is more important than efficiency |
20:26 |
APNG |
just look at V8 if you want an example |
20:26 |
KaadmY |
<APNG> RAM is more important than efficiency |
20:26 |
T4im |
the mod will leak in vanilly certainly, too, just not gonna oom so soon |
20:27 |
KaadmY |
then you complain minetest is laggy? |
20:29 |
APNG |
KaadmY, lookup tables are much faster than CPU heavy shit |
20:29 |
APNG |
lookup tables can speed up minetest significantly |
20:29 |
APNG |
but they're not luajit-friendly |
20:29 |
red-001 |
?? |
20:30 |
APNG |
luajit has a 2GB ram limit |
20:30 |
APNG |
lua doesn't have a ram limit |
20:30 |
APNG |
so in plain lua you better use lookup tables for performance, while in luajit you better use CPU so your code doesn't crash |
20:32 |
red-001 |
2GB is a pretty reasonable limit |
20:32 |
APNG |
some servers have 3TB of RAM >.> |
20:33 |
APNG |
2GB is the opposite of reasonable >.> |
20:35 |
red-001 |
so around $3100 of ram? |
20:36 |
APNG |
RAM is cheap |
20:36 |
APNG |
$3k for 3TB is cheap |
20:37 |
red-001 |
why does lua jit have a limit of 2GB any way? |
20:38 |
T4im |
red, note the #lua channel topic |
20:38 |
* T4im |
starts to get it now |
20:38 |
red-001 |
? |
20:42 |
rubenwardy |
Topic for #lua is: http://lua-users.org/wiki/IrcChannel :: Paste on http://codepad.org/?lang=Lua :: Please consider registering with NickServ if you haven't already. :: Pay no attention to Soni/cat5e/APNG :: * is zero or more :: https://www.youtube.com/watch?v=ubdek0l4Mzc |
20:42 |
APNG |
red-001, something about the way the GC is designed |
20:42 |
|
Calinou joined #minetest-dev |
20:42 |
APNG |
the bytecode isn't quite fully portable |
20:52 |
Fixer |
T4im: last time I checked vanilla used like 2-5 mb of lua mem |
20:52 |
Fixer |
thats nothing |
21:22 |
|
Tmanyo joined #minetest-dev |
21:30 |
|
betterthanyou710 joined #minetest-dev |
21:34 |
|
betterthanyou710 joined #minetest-dev |
21:36 |
|
APNG left #minetest-dev |
21:37 |
|
betterthanyou710 joined #minetest-dev |
22:03 |
|
Grandolf joined #minetest-dev |
22:15 |
hmmmm |
even if you had 3TB of ram, you certainly wouldn't want lua to use it, because of GC churn |
22:19 |
|
betterthanyou710 joined #minetest-dev |
22:30 |
paramat |
the new reused noise 'buffer' must help, i've updated but i'm sure many mods have not |
22:32 |
|
ekem joined #minetest-dev |
22:37 |
|
proller joined #minetest-dev |
22:43 |
Fixer |
darkage probably needs heavy rewriting of mapgen code, since it halts on on_generated |
22:43 |
Fixer |
while adding ores and blocks |
22:53 |
|
halt_ joined #minetest-dev |
22:53 |
red-001 |
what c++ standard should be used when coding for minetest? |
22:55 |
rubenwardy |
red-001, the one before c++11 |
22:56 |
rubenwardy |
C++98? |
22:56 |
rubenwardy |
yeah |
22:57 |
red-001 |
could a newer standard be used? |
22:57 |
red-001 |
or is compiler support not good enough? |
22:58 |
rubenwardy |
I think FreeBSD still has an old version of gcc, or smth |
22:58 |
rubenwardy |
I remember something about switching over when Ubuntu(?) 14.04(?) reaches EOL |
22:59 |
red-001 |
to c++11? |
23:13 |
paramat |
the darkage code looks 'unoptimal' to me |
23:20 |
paramat |
will merge games 1270 1274 in a moment |
23:31 |
|
DI3HARD139 joined #minetest-dev |
23:31 |
rubenwardy |
game#1270 game#1274 |
23:31 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/1270 -- Default, stairs, doors: Vary wood flammable and choppy group values by paramat |
23:31 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/1274 -- eliminated redundant calls by pinkysnowman |
23:32 |
rubenwardy |
Awesome |
23:34 |
paramat |
now merging |
23:39 |
paramat |
complete! |
23:40 |
paramat |
minor change to the last commit, moved the line to a suitable place |
23:44 |
paramat |
#4475 seems good now |
23:44 |
ShadowBot |
https://github.com/minetest/minetest/issues/4475 -- Small improvements of settingtypes.txt by red-001 |
23:47 |
paramat |
rubenwardy i might merge some engine stuff, shall i merge #4416 ? |
23:47 |
ShadowBot |
https://github.com/minetest/minetest/issues/4416 -- Formspecs: Fix background elements from interferring with each other by rubenwardy |
23:48 |
rubenwardy |
Yeah, I think I'll self approve that |
23:48 |
rubenwardy |
thanks |
23:48 |
paramat |
might merge #4372 also |
23:48 |
ShadowBot |
https://github.com/minetest/minetest/issues/4372 -- Disallow registering users with the same name by SmallJoker |
23:48 |
paramat |
ok |
23:49 |
|
Tmanyo joined #minetest-dev |
23:50 |
paramat |
will merge those 3 in a moment |