| Time |
Nick |
Message |
| 00:07 |
|
Zeitgeist_ joined #minetest-dev |
| 00:22 |
|
SoniEx2 joined #minetest-dev |
| 05:35 |
|
celeron55_ joined #minetest-dev |
| 05:40 |
|
cimbakahn joined #minetest-dev |
| 05:43 |
|
VargaD joined #minetest-dev |
| 05:51 |
|
kahrl_ joined #minetest-dev |
| 05:54 |
|
fling_ joined #minetest-dev |
| 05:54 |
|
VargaD_ joined #minetest-dev |
| 05:55 |
|
sofar_ joined #minetest-dev |
| 05:55 |
|
cheapie_ joined #minetest-dev |
| 05:55 |
|
sfan5_ joined #minetest-dev |
| 05:57 |
|
fling_ joined #minetest-dev |
| 06:43 |
|
Wayward_One joined #minetest-dev |
| 06:52 |
|
Supertanker2 joined #minetest-dev |
| 07:20 |
OldCoder |
|
| 07:21 |
OldCoder |
Answer to probably uneducated question is requested: Formspec labels are not stored in the DB, right? Just the data is stored. If you change the labels in a mod, they should change regardless of contents of the DB. Is this so? |
| 07:23 |
OldCoder |
Ah |
| 07:23 |
OldCoder |
They *are* stored in the DB after all |
| 07:29 |
|
nrzkt joined #minetest-dev |
| 07:33 |
|
RealBadAngel joined #minetest-dev |
| 07:38 |
|
Sokomine joined #minetest-dev |
| 07:57 |
|
Krock joined #minetest-dev |
| 08:08 |
|
celeron55 joined #minetest-dev |
| 08:21 |
|
celeron55 joined #minetest-dev |
| 08:35 |
|
Hunterz joined #minetest-dev |
| 08:51 |
|
Wayward_One joined #minetest-dev |
| 08:57 |
|
Lunatrius` joined #minetest-dev |
| 09:33 |
|
Calinou joined #minetest-dev |
| 09:46 |
|
neoascetic joined #minetest-dev |
| 09:51 |
|
neoascetic joined #minetest-dev |
| 09:53 |
neoascetic |
Hi everyone. Someone, please, show me the code where leaves removing occurs after tree digged? |
| 09:53 |
nrzkt |
neoascetic, look at the ABM dedicated to it in lua |
| 09:54 |
|
proller joined #minetest-dev |
| 09:54 |
neoascetic |
I am unable to find it. For what should I search? |
| 09:56 |
neoascetic |
As I understand this should be somewhere in minetest_game, since it minimum development game leaves are not removed |
| 10:02 |
|
kilbith joined #minetest-dev |
| 10:04 |
kilbith |
it's sad to see the git tree so dead these days :( |
| 10:12 |
|
rubenwardy joined #minetest-dev |
| 10:14 |
|
proller joined #minetest-dev |
| 10:16 |
neoascetic |
Oh, found it, it is ABM for group:leafdecay... |
| 10:17 |
|
neoascetic left #minetest-dev |
| 10:20 |
|
celeron55 joined #minetest-dev |
| 10:30 |
|
celeron55 joined #minetest-dev |
| 10:45 |
|
celeron55 joined #minetest-dev |
| 10:51 |
|
alket joined #minetest-dev |
| 10:55 |
|
celeron55 joined #minetest-dev |
| 10:56 |
|
Player2 joined #minetest-dev |
| 10:57 |
|
CraigyDavi joined #minetest-dev |
| 11:06 |
|
Gael-de-Sailly joined #minetest-dev |
| 11:30 |
|
Krock joined #minetest-dev |
| 11:41 |
|
DFeniks joined #minetest-dev |
| 12:06 |
|
who_wants_some joined #minetest-dev |
| 12:20 |
|
Terusthebird joined #minetest-dev |
| 12:27 |
|
alket joined #minetest-dev |
| 12:33 |
|
Fixer joined #minetest-dev |
| 12:35 |
|
Terusthebird joined #minetest-dev |
| 12:48 |
|
Fixer joined #minetest-dev |
| 12:51 |
|
damiel joined #minetest-dev |
| 12:53 |
|
Fixer joined #minetest-dev |
| 13:11 |
|
alket joined #minetest-dev |
| 13:19 |
|
paramat joined #minetest-dev |
| 13:20 |
paramat |
PilzAdam when you're around see https://github.com/minetest/minetest/pull/3402#issuecomment-160294849 |
| 13:23 |
|
est31 joined #minetest-dev |
| 13:47 |
|
paramat left #minetest-dev |
| 13:47 |
|
Darcidride joined #minetest-dev |
| 13:47 |
|
est31 joined #minetest-dev |
| 14:01 |
|
zat joined #minetest-dev |
| 14:03 |
|
Icedream joined #minetest-dev |
| 14:54 |
|
zupoman joined #minetest-dev |
| 14:59 |
|
ElectronLibre joined #minetest-dev |
| 15:08 |
|
alket joined #minetest-dev |
| 15:11 |
|
H-H-H joined #minetest-dev |
| 15:15 |
|
hmmmm joined #minetest-dev |
| 15:15 |
|
turtleman_ joined #minetest-dev |
| 15:32 |
|
paramat joined #minetest-dev |
| 15:33 |
est31 |
paramat, the bug you pointed out was only about a new bug, right? |
| 15:33 |
est31 |
but basically the thing PilzAdam fixed in that PR is ok? |
| 15:34 |
paramat |
correct :) |
| 15:34 |
est31 |
soo... merge? |
| 15:36 |
paramat |
hm possibly, i thought it might be better to add the extra fix to that PR, since it' sso closely related, less work for PA |
| 15:36 |
|
alket joined #minetest-dev |
| 15:36 |
est31 |
okay, lets wait a day then |
| 15:37 |
est31 |
if he fixes it in the pr, we merge it before the day ends, if he doesnt fix it in the pr, we merge it once the day is over, and you make a new issue |
| 15:38 |
paramat |
yes, if PA is busy then we can merge it to avoid conflicts |
| 15:40 |
paramat |
i might merge this fairly trivial fix later https://github.com/minetest/minetest/pull/3406 'Dungeongen: Also preserve river water nodes'. highlighting hmmmm just in case you want to look at it |
| 15:43 |
paramat |
soon i might finally add the docs for biome definitions |
| 16:00 |
|
paramat left #minetest-dev |
| 16:12 |
|
Player2 joined #minetest-dev |
| 17:42 |
VanessaE_ |
idea: give each user on a minetest server a unique key (something simple, an ssl hash or something) unique to them - not associated with username or password, and persistent across sessions. |
| 17:42 |
VanessaE_ |
something that can be used by a mod to ban a single user in a multi-user household. |
| 17:43 |
VanessaE_ |
(I presume the key would be generated and stored by the client and sent to the server along with regular login credentials) |
| 17:44 |
|
H-H-H joined #minetest-dev |
| 17:54 |
Krock |
VanessaE_, by using the mac address or what? |
| 17:54 |
VanessaE_ |
Krock: unique, random key |
| 17:54 |
VanessaE_ |
once generated, it's stored and used forever (until deleted) |
| 17:55 |
Krock |
but that cache could be cleared somehow |
| 17:55 |
Krock |
because it'd be requierd to save it on the client side too |
| 17:55 |
VanessaE_ |
yes, unless the key is stored somewhere that the user wouldn't want to delete |
| 17:56 |
VanessaE_ |
I know some multiplayer games use a public/private key sort of setup like ssh |
| 17:56 |
VanessaE_ |
but that's overkill |
| 17:56 |
Krock |
it's always possible to take out a value of a file that's not compiled or encrypted in some way |
| 17:56 |
VanessaE_ |
I know. |
| 17:56 |
VanessaE_ |
I'm just saying it could be made "inconvenient" |
| 17:56 |
Krock |
yup |
| 17:59 |
|
ElectronLibre joined #minetest-dev |
| 18:00 |
DFeniks |
what if i play from multiple clients ? (not at the same time ofc.) |
| 18:01 |
DFeniks |
how its different from banning username? |
| 18:01 |
Krock |
system-wise path... |
| 18:02 |
Krock |
it would be possible to a degree to ban a person |
| 18:02 |
Krock |
not just an username |
| 18:02 |
Krock |
s/wise/wide/ |
| 18:02 |
DFeniks |
^dual boot OS , windows linux |
| 18:03 |
DFeniks |
i really play from both |
| 18:05 |
DFeniks |
and how this would not ban other user from same house? nevermind i got it , you mean if they use different pc |
| 18:05 |
VanessaE_ |
DFeniks: that's correct |
| 18:06 |
VanessaE_ |
as for playing from multiple clients, well, the key in question would need to be stored somewhere separate from the client itself |
| 18:06 |
VanessaE_ |
(e.g. user's home dir. ~/.local/minetest or something) |
| 18:08 |
DFeniks |
as long as i still can have client in one folder , like compiled with run in place |
| 18:08 |
VanessaE_ |
I mean separate from that, too |
| 18:09 |
VanessaE_ |
pick some path 99.9% of linux installs will have (such as ~/.config or ~/.local) and use it |
| 18:09 |
VanessaE_ |
let the client stay wherever |
| 18:09 |
VanessaE_ |
as long as the key doesn't move around |
| 18:12 |
est31 |
we could make it like minecraft and require money to sign up to public servers |
| 18:12 |
DFeniks |
yeah i know like blender for example can be run from folder but settings are stored somewhere in home like ./blender (i dont really remember) |
| 18:12 |
est31 |
this way we could get some money to run them :) |
| 18:12 |
VanessaE_ |
est31: um, no. |
| 18:12 |
est31 |
but otherwise I dont think such a key will be effective |
| 18:12 |
est31 |
it will just lead to an arms race |
| 18:12 |
VanessaE_ |
how so? |
| 18:13 |
DFeniks |
isnt it easy to found and delete that folder thought? |
| 18:13 |
est31 |
its basically security through obscurity |
| 18:13 |
VanessaE_ |
well |
| 18:13 |
est31 |
not good, concept wise |
| 18:13 |
VanessaE_ |
what other way is there, then? |
| 18:13 |
est31 |
ip address, like now. |
| 18:13 |
VanessaE_ |
nope |
| 18:13 |
VanessaE_ |
that doesn't address the concern I was talking about |
| 18:14 |
est31 |
ah I see |
| 18:14 |
est31 |
multi user household... |
| 18:14 |
VanessaE_ |
yeah |
| 18:14 |
VanessaE_ |
this is becoming a problem |
| 18:14 |
est31 |
well, what about whitelisting usernames? |
| 18:14 |
VanessaE_ |
wouldn't work on a public server. |
| 18:15 |
est31 |
I mean you ban the ip, and if somebody complains that their ip got banned because their little sister did grief, then they can get whitelisted |
| 18:15 |
est31 |
doable with our current mod system already. |
| 18:15 |
VanessaE_ |
that could work |
| 18:15 |
VanessaE_ |
I still think a per-user key would be better though |
| 18:17 |
DFeniks |
that might work but maybe not against most dedicated offenders |
| 18:17 |
VanessaE_ |
the idea is to make it reasonably cumbersome for one user to masquerade as another; more traditional methods like ban-by-ip, or dropping them at the firewall, can be used for people who know how to get around that. |
| 18:23 |
rubenwardy |
est32, VanessaE_: most people I ban are just immature kids that know little about computers. They just try a different user name each day, until their ip changes and they get in |
| 18:23 |
rubenwardy |
well |
| 18:23 |
VanessaE_ |
exactly my point |
| 18:23 |
rubenwardy |
from their interactions I guess they are |
| 18:24 |
VanessaE_ |
ban by IP is not sufficient in any case. |
| 18:24 |
VanessaE_ |
/xban is better than /ban but even THAT isn't enough in some situations |
| 18:24 |
rubenwardy |
xban2 is awesome |
| 18:25 |
Fixer |
my ISP moved to crap department, everybody has CGNAT :( So many people under 1 IP |
| 18:25 |
est31 |
rubenwardy, do you know whether they use mobile clients |
| 18:25 |
est31 |
or do they use non mobile ones |
| 18:25 |
VanessaE_ |
of course using a key has an extra benefit: no need for passwords :) |
| 18:25 |
rubenwardy |
I haven't installed any stats |
| 18:26 |
rubenwardy |
Is it possible to make a stats mod to see the OS/device of people connecting? |
| 18:26 |
|
alket joined #minetest-dev |
| 18:26 |
est31 |
rubenwardy, only with a debug build, unfortunately |
| 18:26 |
est31 |
but OS and device are not possible |
| 18:26 |
VanessaE_ |
rubenwardy: or, run a media server for your minetest server. |
| 18:26 |
rubenwardy |
the client has to be though |
| 18:26 |
VanessaE_ |
and get your stats from that. |
| 18:26 |
rubenwardy |
that's good |
| 18:26 |
est31 |
there is no user agent like stuff for minetest, unfortunately |
| 18:26 |
VanessaE_ |
surely the user agent varies between andr... |
| 18:26 |
VanessaE_ |
damn. |
| 18:27 |
est31 |
but yeah, media server will work |
| 18:27 |
rubenwardy |
what's a light weight media server which will record these things? |
| 18:27 |
VanessaE_ |
nginx |
| 18:28 |
VanessaE_ |
scrape the logs |
| 18:30 |
Fixer |
https://stackoverflow.com/questions/3652271/getting-os-version ? |
| 18:30 |
rubenwardy |
not for clients |
| 18:31 |
rubenwardy |
mods run server side |
| 18:31 |
Fixer |
oh :S |
| 18:31 |
Fixer |
damned |
| 18:36 |
Fixer |
for example, CS 1.6 anticheat utilizes cookies to ban efficiently (if you block port 80 in firewall, you can avoid the ban) |
| 18:36 |
Fixer |
but it is somewhat crude and not safe |
| 18:36 |
Fixer |
better avoid the web stuff in mt |
| 18:37 |
est31 |
CS what? |
| 18:37 |
rubenwardy |
counterstrike |
| 18:37 |
est31 |
aha |
| 18:38 |
est31 |
well blocking port 80 is one option |
| 18:38 |
est31 |
or just deleting cookies is the other |
| 18:38 |
Fixer |
nevermind, don't implement web browser in mt, it will open huge security holes :( |
| 18:38 |
est31 |
I doubt we should get supercookies like VanessaE_ promotes |
| 18:38 |
est31 |
they can be abused like normal cookies too |
| 18:38 |
VanessaE_ |
using port 80 isn't really an option anyway; how would it work with several servers on the same machine? |
| 18:38 |
est31 |
VanessaE_ is a nice person, and most of the server owners are too |
| 18:39 |
est31 |
but things can change, cookies werent invented with ad tracking in mind either |
| 18:39 |
rubenwardy |
There's no griefing in CTF, only sabotage - no real loss or destruction. |
| 18:39 |
est31 |
most of the server owners == all that I know |
| 18:40 |
rubenwardy |
if you were to make supercookies, they'd have to be only for banning. Possibly for linking multi-account users |
| 18:43 |
Calinou |
<VanessaE_> idea: give each user on a minetest server a unique key (something simple, an ssl hash or something) unique to them - not associated with username or password, and persistent across sessions. |
| 18:43 |
Calinou |
this is what Quake 3 did with GUIDs, easily fakeable and abusable |
| 18:43 |
Calinou |
it almost became a meme, it was so easy to change GUIDs :p |
| 18:43 |
VanessaE_ |
heh |
| 18:44 |
VanessaE_ |
remember, perfect is the enemy of good :P |
| 18:44 |
est31 |
lets add DRM to minetest |
| 18:44 |
est31 |
and make it closed source |
| 18:44 |
est31 |
then nobody can use cracked minetest clients anymore! |
| 18:44 |
VanessaE_ |
:P |
| 18:44 |
RealBadAngel |
lol |
| 18:44 |
Calinou |
est31, let's reinvent PunkBuster! |
| 18:44 |
Calinou |
admin-requested clients of screenshots. |
| 18:44 |
Calinou |
100% cheatproof |
| 18:44 |
Calinou |
screenshots of clients* |
| 18:45 |
RealBadAngel |
est31, wanna feature request? |
| 18:45 |
VanessaE_ |
you guys joke about it, but I'm being serious. |
| 18:46 |
VanessaE_ |
it's a problem. there's gotta be some kind of solution that works well enough. |
| 18:46 |
RealBadAngel |
quite easy, all the code needed is here |
| 18:46 |
rubenwardy |
Just because it's not fool proof doesn't mean it's not needed |
| 18:46 |
Calinou |
yes |
| 18:46 |
Calinou |
it is possible to have a GUID-like system, Xonotic does that for its stats system. |
| 18:46 |
Calinou |
it is useful. |
| 18:46 |
rubenwardy |
However, I don't know if I like the idea |
| 18:46 |
Calinou |
but don't think it's reliable for banning |
| 18:47 |
RealBadAngel |
est31, do you dare or not? |
| 18:48 |
Fixer |
please, no pankbaster |
| 18:49 |
RealBadAngel |
chicken :) |
| 18:52 |
RealBadAngel |
but seriously, somebody please implement Lua join(rejoin) to the server |
| 18:52 |
rubenwardy |
What do you mean? |
| 18:52 |
rubenwardy |
redirect, or reconnect? |
| 18:52 |
Fixer |
yes, without even unloading everything |
| 18:52 |
Fixer |
reconnect |
| 18:52 |
RealBadAngel |
merge all the servers |
| 18:53 |
VanessaE_ |
you mean like linking books in MC? |
| 18:53 |
RealBadAngel |
by walkin thru gate or somethin like that |
| 18:53 |
RealBadAngel |
i mean literally merge all the servers into one big web |
| 18:54 |
RealBadAngel |
thx to the connections between them |
| 18:54 |
VanessaE_ |
right |
| 18:54 |
Fixer |
you connect to server it has big tablo with signs that has server info, you punch some and connect to it |
| 18:54 |
RealBadAngel |
how to enter the gate is not a problem |
| 18:55 |
VanessaE_ |
I guess I meant portals anyway |
| 18:55 |
VanessaE_ |
linking book is some other game |
| 18:55 |
Fixer |
yes |
| 18:55 |
RealBadAngel |
we need reconnect (and failback) funtions |
| 18:55 |
VanessaE_ |
it's a good idea |
| 18:55 |
RealBadAngel |
we can define rules later on |
| 18:55 |
RealBadAngel |
like what player can carry |
| 18:56 |
VanessaE_ |
by default, nothing |
| 18:56 |
VanessaE_ |
(I would hope) |
| 18:56 |
RealBadAngel |
we can have rich mine worlds, just building ones |
| 18:56 |
RealBadAngel |
you would like to visit rich world |
| 18:57 |
RealBadAngel |
just to dig a lot |
| 18:57 |
Fixer |
we heading in wrong direction |
| 18:57 |
Fixer |
are |
| 18:58 |
RealBadAngel |
with this idea? |
| 18:58 |
Fixer |
we need celeron55 finished farmap to unblock the development |
| 18:58 |
Calinou |
farmap is overrated. |
| 18:58 |
Calinou |
just optimize generic block rendering further |
| 18:58 |
VanessaE_ |
farmap is useful but other stuff needs done with it |
| 18:58 |
Calinou |
then players can increase their view distance to match's Minecraft |
| 18:58 |
VanessaE_ |
like that ^^^ |
| 18:58 |
Calinou |
(which is pretty far by default) |
| 18:58 |
RealBadAngel |
farmesh is something to improve your view |
| 18:58 |
Calinou |
I remember when I played Minecraft in 2010, I was amazed by the default "far" view distance |
| 18:58 |
Fixer |
yes, but celeron55 asked no touch that stuff until farmap is ready, and farmap branch is 14 days old now |
| 18:59 |
VanessaE_ |
if celeron55 isn't going to finish farmap soon, then the other rendering code needs improved. |
| 18:59 |
RealBadAngel |
connecting all the worlds is something different |
| 18:59 |
VanessaE_ |
RealBadAngel: I think he means that distance rendering is needed so that a given server appears sufficiently far into the distance? |
| 18:59 |
RealBadAngel |
its gonna affect the gameplay and community |
| 18:59 |
Fixer |
celeron55, hi, any news on farmap, when will it be finished? |
| 18:59 |
RealBadAngel |
not just what one folk can see on the monitor |
| 18:59 |
Fixer |
RealBadAngel, no it is not |
| 19:00 |
Fixer |
RealBadAngel, it is just ingame server browser |
| 19:00 |
|
JohnnyComeL8ly joined #minetest-dev |
| 19:00 |
Fixer |
RealBadAngel, we have lots of server but few players actually |
| 19:00 |
RealBadAngel |
fixer, u got it wrong, totally |
| 19:00 |
RealBadAngel |
i meant INGAME |
| 19:00 |
Calinou |
could we experiment with defaulting to dedicated_server_step = 0.05? |
| 19:00 |
Calinou |
it could result in more responsive gameplay |
| 19:00 |
Calinou |
I've set that on my Hunger Games server |
| 19:00 |
Fixer |
RealBadAngel, like in MC? |
| 19:01 |
RealBadAngel |
Calinou, nrzkt set it to 0.005 |
| 19:01 |
RealBadAngel |
0.025 |
| 19:01 |
Calinou |
well if you have spare CPU, that can be OK |
| 19:01 |
Calinou |
but why not default to 0.05? my single-core server does well with that |
| 19:01 |
RealBadAngel |
Fixer,mc doesnt have connected world |
| 19:01 |
VanessaE_ |
Calinou: throw playlife + biome_lib at it. |
| 19:01 |
rubenwardy |
What is it that limits Minetest to ~20 players before lag? |
| 19:01 |
VanessaE_ |
that's a good way (imho) to test globalstep speeds. |
| 19:02 |
RealBadAngel |
i mean smooth transitons from one server to another |
| 19:02 |
Calinou |
VanessaE_, maybe mods could apply a factor to the server step |
| 19:02 |
rubenwardy |
plantlife? |
| 19:02 |
RealBadAngel |
without exiting to the menu |
| 19:02 |
Calinou |
so that, as soon as heavy mods are installed, the server step is doubled (halved FPS) |
| 19:02 |
RealBadAngel |
just stepping the gate |
| 19:02 |
VanessaE_ |
rubenwardy: I've had as many as 40 on one of my servers without significant lag. it's all in the network performance really |
| 19:02 |
VanessaE_ |
rubenwardy: yeah, because biome_lib by itself doesn't do anything useful. |
| 19:02 |
Fixer |
RealBadAngel, that is sooo far ahead, you should better polish PR 3166 that will bring hope to many |
| 19:03 |
RealBadAngel |
Fixer, damn, this idea is far simpler to make |
| 19:03 |
Fixer |
RealBadAngel, and occlusion culing per block |
| 19:03 |
RealBadAngel |
lol |
| 19:03 |
rubenwardy |
VanessaE_, you said playlife |
| 19:03 |
RealBadAngel |
and what? fix irrlicht? |
| 19:03 |
RealBadAngel |
:) |
| 19:03 |
VanessaE_ |
rubenwardy: shit. :) |
| 19:03 |
Fixer |
yes!11 |
| 19:04 |
VanessaE_ |
RealBadAngel: "Stay on target. Stay on target!" |
| 19:04 |
Calinou |
would it possible to do occlusion culling on GPU? |
| 19:04 |
RealBadAngel |
i do have my own targets |
| 19:04 |
Calinou |
using hardware occlusion queries |
| 19:04 |
RealBadAngel |
but YOU dont have any |
| 19:04 |
Fixer |
RealBadAngel, occlusion is working on per 8*8*8 blocks basis, somewhere on wiki i've seen it needs to be made per each block to improve (maybe celeron wrote it idn) |
| 19:05 |
VanessaE_ |
Fixer: it needs to be done on a per-NODE basis I think. |
| 19:05 |
Fixer |
yes |
| 19:05 |
VanessaE_ |
let the damn GPU do its job |
| 19:05 |
Fixer |
per node |
| 19:05 |
RealBadAngel |
fixer, im coding candies, rarely something other :P |
| 19:05 |
Fixer |
RealBadAngel, candies slow down everything |
| 19:05 |
RealBadAngel |
ofc |
| 19:05 |
RealBadAngel |
there are no free candies :P |
| 19:05 |
Fixer |
smth is wrong with MT |
| 19:05 |
Fixer |
too much lags |
| 19:06 |
Fixer |
everybody is complaining |
| 19:06 |
Fixer |
it is an issue |
| 19:06 |
Calinou |
per-node is probably too much, even if on the GPU |
| 19:06 |
RealBadAngel |
3166 is something about it |
| 19:06 |
Fixer |
lag or low fps |
| 19:06 |
Calinou |
Fixer, yeah, even my new laptop doesn't run Minetest that fast |
| 19:06 |
VanessaE_ |
Calinou: per mapblock then |
| 19:06 |
RealBadAngel |
but nobody is willing to help me with that |
| 19:06 |
Calinou |
there should be a release dedicated to otpimization… |
| 19:06 |
Calinou |
both server and client |
| 19:06 |
VanessaE_ |
anyway something other than some huge quantity |
| 19:07 |
VanessaE_ |
Calinou: 0.4.14 was supposed to be bugfix + optimization release. |
| 19:07 |
Calinou |
VanessaE_, well for a start, occlusion culling could be toggleable on client |
| 19:07 |
VanessaE_ |
that was the plan but I guess that got thrown out already |
| 19:07 |
Calinou |
so that we can see if it actually helps (it might not) |
| 19:07 |
Fixer |
it is not even 0.4.13.1 right now :} |
| 19:07 |
Fixer |
i lied somewhat |
| 19:08 |
rubenwardy |
#3166 |
| 19:08 |
VanessaE_ |
https://github.com/minetest/minetest/pull/3166 |
| 19:08 |
VanessaE_ |
Send to clients only changed node metadata instad of whole mapblock |
| 19:08 |
rubenwardy |
ah |
| 19:09 |
rubenwardy |
thats still not done? |
| 19:09 |
VanessaE_ |
nope. RealBadAngel needs help with it. |
| 19:09 |
Fixer |
RealBadAngel, let me view dozens of programming lectures, read python book, c++ book, lua book, and maybe I will help somehow (in this century)... or not :p accountant is too damb for that (actually not) |
| 19:10 |
Calinou |
I'd like Minetest to get a fancy UI |
| 19:10 |
rubenwardy |
samn |
| 19:10 |
Calinou |
Doomsday has awesome UI, made using Qt 5 |
| 19:10 |
rubenwardy |
* same |
| 19:10 |
Fixer |
please fix bugs, make fancy later |
| 19:11 |
Fixer |
it is slow as snail |
| 19:11 |
rubenwardy |
true |
| 19:11 |
Fixer |
it is very slow |
| 19:11 |
rubenwardy |
but part of fixing formspecs and HUD is making the UI better |
| 19:11 |
Fixer |
MC can handle 100 of players |
| 19:11 |
Fixer |
MT becames slow with >10 players |
| 19:11 |
exio4 |
\o/ |
| 19:11 |
rubenwardy |
hudkit mades hud bearable |
| 19:11 |
VanessaE |
MC has the advantage that it spreads the load too |
| 19:11 |
rubenwardy |
but still not |
| 19:11 |
rubenwardy |
good |
| 19:11 |
exio4 |
this works |
| 19:12 |
VanessaE |
I've been told that the CPU usage on any given machine or server is all spread out across all connected machines, like some sort of basic clustering. |
| 19:12 |
exio4 |
Calinou: you were the one saying "boost is nice, wth are we not using C++11?" right |
| 19:12 |
exio4 |
VanessaE: er, what, that must be really new |
| 19:12 |
Fixer |
when someone goes to explore in multiplayer server starts to lag for everybody |
| 19:12 |
VanessaE |
exio4: my info may be wrong. it's just something I recall reading somewhere. |
| 19:12 |
exio4 |
as far as I know the last stable release of minecraft doesn't do that |
| 19:12 |
exio4 |
latest* |
| 19:13 |
VanessaE |
hm. wonder where I got that from then |
| 19:13 |
exio4 |
minetest should be rewritten |
| 19:13 |
exio4 |
it is a lot of work |
| 19:14 |
exio4 |
but people have been fighting with the current code for like 2 years |
| 19:14 |
exio4 |
dont they realize it's insane to keep going on that? |
| 19:14 |
VanessaE |
who's gonna rewrite it? |
| 19:14 |
VanessaE |
that's not just a "lot" of work, it's an insane amount. |
| 19:15 |
exio4 |
let's see how much we can get with donations and hire some of the current devs? |
| 19:15 |
VanessaE |
that's been done. the answer is "nowhere". |
| 19:15 |
exio4 |
i didnt know about that one |
| 19:15 |
Calinou |
http://archive.hugo.pro/doomsday-ui.webm |
| 19:15 |
exio4 |
VanessaE: let's consider this though |
| 19:16 |
Calinou |
(this is Doomsday Engine) |
| 19:16 |
exio4 |
VanessaE: how `much` work is it to fix the current bugs? |
| 19:16 |
Calinou |
it likes to use fancy transitions for sure :p |
| 19:16 |
Fixer |
better hope that irrlicht will not die some day... it is already in a hospital |
| 19:16 |
VanessaE |
exio4: I couldn't give you a figure, but "a fuckton" might be close. |
| 19:16 |
Calinou |
Fixer, Urho3D could be used for a rewrite :P |
| 19:16 |
exio4 |
VanessaE: so you're comparing two insane-level amounts of work |
| 19:17 |
VanessaE |
Calinou: wasn't c55 targeting that for his buildat thing? |
| 19:17 |
Fixer |
Calinou, i will start the rewrite, will be finished close to "heat death of the universe" |
| 19:18 |
VanessaE |
exio4: the way I've seen people talk around here, fixing all of the bugs would be more work than a total rewrite, which would be more work than rewriting, let's say, The GIMP. :P |
| 19:18 |
Fixer |
and now Zeno has quit and we stuck on irrlicht |
| 19:18 |
VanessaE |
I think the real problem is just time. |
| 19:18 |
VanessaE |
no one has any |
| 19:18 |
exio4 |
time = money |
| 19:18 |
Fixer |
yes |
| 19:18 |
exio4 |
we thus, need money |
| 19:18 |
VanessaE |
exio4: fuck that |
| 19:19 |
VanessaE |
I refuse to believe that |
| 19:19 |
Fixer |
development is so slow here, look how many PRs out there |
| 19:19 |
Calinou |
VanessaE, yes, he used Urho3D for it |
| 19:19 |
Calinou |
it's pretty decent IMO |
| 19:19 |
VanessaE |
time = money if, and only if, you would otherwise spend that time making money at your job or something. |
| 19:19 |
Calinou |
comes with AngelScript, or Lua scripting |
| 19:19 |
|
behalebabo joined #minetest-dev |
| 19:20 |
VanessaE |
if that time would otherwise be spent in some leisure activity, it does NOT euqate to money! |
| 19:20 |
VanessaE |
equate* |
| 19:20 |
Fixer |
people has very few time to review PRs |
| 19:20 |
rubenwardy |
If you paid the devs so they could work on MT, instead of their (old) job |
| 19:21 |
exio4 |
they'd need to accept a not-insane salary though :P |
| 19:21 |
VanessaE |
define "not insane" |
| 19:21 |
Fixer |
like few dollars per month (from donations) |
| 19:21 |
Fixer |
$) |
| 19:21 |
VanessaE |
seems the going rate for a good coder is around $150k a year now |
| 19:21 |
Calinou |
$150k? wow |
| 19:21 |
exio4 |
nah, the idea is to first get the donations, then see how much we get |
| 19:21 |
exio4 |
VanessaE: that's an american-salary |
| 19:21 |
VanessaE |
50-60k/year if you're just "so so" |
| 19:21 |
RealBadAngel |
what are you talkin about? |
| 19:21 |
Calinou |
yeah, usually developers refuse getting less than $50k |
| 19:22 |
VanessaE |
Calinou: yeah, it's ridiculous. |
| 19:22 |
exio4 |
VanessaE: which, compared to pretty much every other country on earth, is insanely high |
| 19:22 |
RealBadAngel |
you mean all those yrs of coding stuff was for money? |
| 19:22 |
rubenwardy |
ofc |
| 19:22 |
Fixer |
developers got very good salaries in East Europe, others just work for food *cough* |
| 19:22 |
RealBadAngel |
i do work just for food |
| 19:22 |
VanessaE |
RealBadAngel: an old, stupid americanism. "Time is money." |
| 19:23 |
rubenwardy |
You could make it enough for a part time job, alongside another one |
| 19:23 |
RealBadAngel |
but i dont care |
| 19:23 |
Fixer |
VanessaE, but it works |
| 19:23 |
Fixer |
VanessaE, runs the economy |
| 19:23 |
VanessaE |
Fixer: it works if you're a greedy bastard :P |
| 19:23 |
RealBadAngel |
my "real" life is coding |
| 19:23 |
VanessaE |
"YOu mean I've only been a capitalist for four hours and I already owe $10,000 ??" "That's what makes our system work -- everybody owes everybody." |
| 19:24 |
VanessaE |
(might have botched the quote. from the movie "One, Two, Three") |
| 19:24 |
exio4 |
it's easy, if I want to live with certain standard of living, you need money, one way to get money legally is to work, you thus decide to work, and you spend time on something, for a monetary recompensation |
| 19:24 |
RealBadAngel |
its just a point in my timeline when i gen fire up browser, check the forums |
| 19:24 |
VanessaE |
exio4: yes, TO A POINT. |
| 19:24 |
RealBadAngel |
see what was done |
| 19:24 |
exio4 |
VanessaE: capitalism is what the US has given to the world! you should be proud of it! |
| 19:24 |
RealBadAngel |
and see if i can do anything more |
| 19:24 |
VanessaE |
there's a point beyond which working for money is useless (even in the face of the extra income) |
| 19:25 |
exio4 |
yes, obviously |
| 19:25 |
exio4 |
you won't hire somebody at random |
| 19:26 |
RealBadAngel |
hiring somebody to do hobby work is kinda pointless |
| 19:26 |
RealBadAngel |
you will offecnce all the ones doing it for free |
| 19:27 |
exio4 |
whatever, I wouldn't feel offended if somebody is being paid to do the nasty work =) |
| 19:28 |
RealBadAngel |
nasty work could be made if folk start to cooperate |
| 19:28 |
RealBadAngel |
no need for extra payment |
| 19:28 |
RealBadAngel |
just a simple kick in the ass ;) |
| 19:28 |
VanessaE |
I make my stuff for free because I want to make stuff people will find useful. Good legacy, I think. |
| 19:29 |
Fixer |
exio4, it has its faults, but it works, don't ask me about communist regimes, my country lost millions of people because of it (capitalism with some state regulation is good, law and order must be supported though). |
| 19:29 |
exio4 |
VanessaE: how do you spend your time? |
| 19:29 |
VanessaE |
(even when I sell hardware, others are welcome to make those things themselves too) |
| 19:29 |
VanessaE |
exio4: most of the time? reading or watching movies. |
| 19:29 |
RealBadAngel |
Fixer, whats ur nationality btw? |
| 19:29 |
Fixer |
ukrainian |
| 19:29 |
exio4 |
VanessaE: ok, I mean, how do you get your money? |
| 19:29 |
RealBadAngel |
pretty close, im polish |
| 19:30 |
VanessaE |
exio4: I am disabled. |
| 19:30 |
VanessaE |
(so, federal money) |
| 19:30 |
Fixer |
just avoid totalitarian leftist regimes... it sucked too much blood from people |
| 19:30 |
RealBadAngel |
fixer, do you know we were one country once? |
| 19:30 |
VanessaE |
when I worked, it varied. labor of some kind usually. customer service. did sysadmin work for a while too (not that good at it though, at the time). |
| 19:31 |
Fixer |
RealBadAngel, we all changed countries a lot of times, our nation is finally free and independent from Ussr/Russia |
| 19:31 |
Fixer |
but it is offtopic |
| 19:31 |
exio4 |
VanessaE: ok, that explains your position, nvm |
| 19:31 |
VanessaE |
JohnnyComeL8ly: it was. now it's #capitalist-bitch-session |
| 19:31 |
exio4 |
Fixer: any extreme is bad |
| 19:31 |
VanessaE |
exio4: when I was working, the same rule applied. |
| 19:32 |
exio4 |
(iow, the US and the USSR are bad!) |
| 19:33 |
Fixer |
i don't want to live in that |
| 19:33 |
VanessaE |
Fixer: unless you say the wrong thing to the wrong person. then expect to have your life financially ruined, here. |
| 19:34 |
exio4 |
let's go with a 5-day week, 120 hours, you spend 30 hours working, 10 doing things directly-related to work, 40 hours spent sleeping, leaves you with 40 hours for fun per 5-day week,, assume you have family / friends, reduced to 10/15 hours per week on anything you want |
| 19:34 |
VanessaE |
(jail, not usually so much. but corps love to sue) |
| 19:35 |
Fixer |
VanessaE, that is more USA thing, Europe was always more people oriented |
| 19:35 |
RealBadAngel |
old times are a thing to bitch on |
| 19:35 |
VanessaE |
Fixer: so I assume |
| 19:35 |
RealBadAngel |
now we are independent |
| 19:35 |
Fixer |
Europe is somewhat left (socialistic) |
| 19:35 |
exio4 |
see, this is why I wanted to move the conversation to the main channel VanessaE :p |
| 19:35 |
RealBadAngel |
no matter when we live |
| 19:36 |
rubenwardy |
except England |
| 19:36 |
rubenwardy |
very right wing |
| 19:36 |
RealBadAngel |
now its the time to join the worlds |
| 19:36 |
RealBadAngel |
also in mt |
| 19:36 |
JohnnyComeL8ly |
This is why ppl freak out about -dev being flooded.... :P |
| 19:36 |
Fixer |
sometimes i want more right here, populism in socialism can ruin everything |
| 19:37 |
exio4 |
populism is the worst |
| 19:37 |
rubenwardy |
I'm willing to offer bounties for features, but I only have finite money, and not everyone is motivated by bounties |
| 19:37 |
RealBadAngel |
JohnnyComeL8ly, we are talkin now about the ideas that can change the game |
| 19:37 |
Fixer |
populism is what was in Greece and many exUSSR countries |
| 19:37 |
RealBadAngel |
each server has a few players |
| 19:38 |
RealBadAngel |
we do have 20+ servers |
| 19:38 |
exio4 |
rubenwardy: that sounds nicer |
| 19:38 |
RealBadAngel |
now lets make new player join |
| 19:38 |
RealBadAngel |
he will see 100+ folks |
| 19:38 |
exio4 |
rubenwardy: have some central entity to pick up money and then setup bounties on things :P |
| 19:38 |
RealBadAngel |
on all the servers |
| 19:39 |
RealBadAngel |
snowball effect |
| 19:39 |
Fixer |
example, in Ukraine heating cost was 5 cent/m^3, while real cost was 9x more, difference was paid by state, there was "cheap" heating for people and costly for enterprises. That was proven to be a disaster |
| 19:39 |
Fixer |
state incurret HUGE loss |
| 19:39 |
Fixer |
100 of billion just on heating |
| 19:39 |
Fixer |
that kind of economy will ruin everything |
| 19:40 |
Fixer |
now they are moving to real price for all with subsidies for poor people |
| 19:40 |
Fixer |
before that subsidies were for all |
| 19:41 |
Fixer |
yes, it is development channel, sry, I hope celeron will not ban us this day |
| 19:41 |
RealBadAngel |
JohnnyComeL8ly, case is to move development forward |
| 19:41 |
Fixer |
but please highlight him, he wanted to finish farmap, maybe he waits for some poke :} |
| 19:41 |
RealBadAngel |
candies, farmesh wont get us nowhere |
| 19:41 |
RealBadAngel |
but fresh and unique ideas can |
| 19:42 |
JohnnyComeL8ly |
Yes. |
| 19:42 |
Fixer |
nobody will code it anyway |
| 19:42 |
Fixer |
:} |
| 19:42 |
RealBadAngel |
Fixer, code for that is already in the engine |
| 19:42 |
JohnnyComeL8ly |
I think that the game should be simplified before it is advanced. |
| 19:42 |
RealBadAngel |
you are using it each time you join the server |
| 19:42 |
Fixer |
minetest lags... it add to my depression |
| 19:42 |
Fixer |
adds* |
| 19:43 |
RealBadAngel |
lags, srags or whatever |
| 19:43 |
RealBadAngel |
shit happens |
| 19:43 |
Fixer |
first you need responsiveness boost |
| 19:43 |
RealBadAngel |
its your shit and get used to it |
| 19:43 |
Fixer |
lol, yes |
| 19:43 |
RealBadAngel |
at least you can yell at others |
| 19:44 |
RealBadAngel |
its your sandbox :) |
| 19:44 |
Fixer |
liquid bugs are horrible please fix!!1 |
| 19:44 |
Fixer |
that hanging water columns... |
| 19:44 |
Fixer |
they are on every server |
| 19:44 |
JohnnyComeL8ly |
MT isn't responsible for infrastructure... it is too bad not everyone is able to just connect to the same lan.... |
| 19:44 |
JohnnyComeL8ly |
That would be awesome! |
| 19:44 |
RealBadAngel |
not the same lan |
| 19:45 |
RealBadAngel |
all the servers around and you can WALK to join another |
| 19:45 |
RealBadAngel |
thats the point |
| 19:46 |
JohnnyComeL8ly |
Yeah! |
| 19:46 |
RealBadAngel |
server admins can set up gates, with rules |
| 19:46 |
RealBadAngel |
in no time we can have all connected |
| 19:46 |
Fixer |
look at LinuxGaming2 server, after moving to other VPS and to 0.4.13 it has huge max_lag from time to time, and admin can't mitigate it (mod profilier is too hard for him) |
| 19:47 |
|
est31 joined #minetest-dev |
| 19:47 |
Fixer |
and I get up to 15 fps in stable (thats without technic or mesecons iirc) |
| 19:47 |
Fixer |
in spawn* |
| 19:48 |
Calinou |
some people just use too many mods |
| 19:49 |
VanessaE |
"too many mods" really shouldn't affect FPS |
| 19:49 |
Fixer |
they have only 79 mods that are quite simple |
| 19:49 |
VanessaE |
especially if those mods are on some remote server. |
| 19:50 |
exio4 |
entities have to be rendered by the client |
| 19:50 |
VanessaE |
(the wrong *choice* of mods can, such as that ^^^) |
| 19:50 |
exio4 |
and nodes need to be known by the client too |
| 19:50 |
* Fixer |
.oO(also ... Inventory lag problem!) |
| 19:52 |
est31 |
well yeah VanessaE is right |
| 19:53 |
est31 |
there is only this meta change stuff |
| 19:53 |
est31 |
but that shouldnt affect |
| 19:53 |
* est31 |
hopes we store GUI outside of meta one day |
| 19:53 |
RealBadAngel |
est31, meta stuff is affecting developed servers a lot |
| 19:54 |
RealBadAngel |
more than you can imagine |
| 19:54 |
est31 |
well yeah, this has to be fixed |
| 19:55 |
RealBadAngel |
it can be fixed, with some help |
| 19:55 |
RealBadAngel |
ive made most of the work |
| 19:55 |
VanessaE |
hence #3166... |
| 19:55 |
RealBadAngel |
and i dont mean here whities comments |
| 19:55 |
VanessaE |
I still say merge it now and fix the double-send glitch later. |
| 19:55 |
VanessaE |
2 nodes' worth is better than 7 whole mapblocks' worth. |
| 19:56 |
RealBadAngel |
sesnding logic is quite fixed |
| 19:56 |
est31 |
<VanessaE> I still say merge it now and fix the double-send glitch later. |
| 19:56 |
RealBadAngel |
i dont have an idea how to resolve it |
| 19:56 |
est31 |
you are the one that cries "stability" dont you? |
| 19:57 |
RealBadAngel |
est31, that commit is tested for 2 month already |
| 19:57 |
Fixer |
granted fixing lags is not stability issues -_- |
| 19:57 |
VanessaE |
est31: I am, but I've tested that code in the past. it was stable for me, though it didn't offer my as big an improvement as others may see. |
| 19:57 |
RealBadAngel |
and it fucking workss |
| 19:57 |
VanessaE |
offer me* |
| 19:57 |
RealBadAngel |
its a part of epixel |
| 19:57 |
est31 |
nrzkt's changes were "tested for 2 months" too, and they introduced tons of regressions either way |
| 19:58 |
VanessaE |
est31: fair enough. |
| 19:58 |
RealBadAngel |
maybe |
| 19:58 |
VanessaE |
but we're gonna get nowhere if no one tests and helps complete the code |
| 19:58 |
RealBadAngel |
but that one reduces load a lot |
| 19:58 |
RealBadAngel |
think for chirst sake |
| 19:58 |
est31 |
RealBadAngel, just fix the bug, its so trivial, and you still dont fix it |
| 19:58 |
Fixer |
can developers focus on helping RBA with 3166? |
| 19:58 |
RealBadAngel |
bug? |
| 19:58 |
VanessaE |
RealBadAngel: the double-send glitch? |
| 19:59 |
Fixer |
you should be a team |
| 19:59 |
est31 |
and the send-to-completely-remote-clients bug |
| 19:59 |
est31 |
which is far worse |
| 19:59 |
VanessaE |
oh yeah, I forgot about that one |
| 19:59 |
RealBadAngel |
double send glitch is lack of queue |
| 19:59 |
est31 |
you stand at spawn, get the metadata changes for the whole fucking server |
| 19:59 |
est31 |
double send isnt that much of an issue |
| 19:59 |
RealBadAngel |
est31, ah i see |
| 20:00 |
VanessaE |
est31: is that a regression? |
| 20:00 |
est31 |
VanessaE, yes. |
| 20:00 |
VanessaE |
eepo |
| 20:00 |
RealBadAngel |
that can be resolved in the same queue |
| 20:00 |
VanessaE |
eep* |
| 20:02 |
RealBadAngel |
damn, i just cant figure how to make that queue properly |
| 20:02 |
RealBadAngel |
this is where im stuck |
| 20:02 |
RealBadAngel |
dont u get it? |
| 20:02 |
est31 |
hrmmm |
| 20:03 |
est31 |
its basically just a bitset, not a queue |
| 20:03 |
est31 |
really huge bitset |
| 20:03 |
RealBadAngel |
its tick time |
| 20:03 |
est31 |
can be done via hashmap I guess. |
| 20:04 |
RealBadAngel |
collect all the meta changes, remove duplicates, and send them as list |
| 20:04 |
RealBadAngel |
farplayers can be used to remove clients that should not receive the update |
| 20:05 |
RealBadAngel |
and just get setBlockNotSent instead |
| 20:05 |
est31 |
exactly |
| 20:05 |
est31 |
so do it, no queue needed |
| 20:05 |
RealBadAngel |
atm 3166 has a single meta change send |
| 20:05 |
RealBadAngel |
but it can be a list |
| 20:05 |
RealBadAngel |
theres code to process lists |
| 20:06 |
RealBadAngel |
so we can bunch them |
| 20:08 |
RealBadAngel |
theres a loop already, to process all the events, map update ones |
| 20:08 |
RealBadAngel |
far players is declared just before |
| 20:09 |
RealBadAngel |
i feel like some of the logic should be made inside this very loop |
| 20:09 |
RealBadAngel |
to send blockNotSent |
| 20:10 |
RealBadAngel |
and the rest oustide |
| 20:10 |
RealBadAngel |
this way i can get rid of far players |
| 20:10 |
RealBadAngel |
and maintain compability with older clients at the same time |
| 20:18 |
VanessaE |
so what's gonna be? |
| 20:18 |
VanessaE |
hem and haw over 3166 or fix it? |
| 20:23 |
|
Obani joined #minetest-dev |
| 20:23 |
celeron55 |
stability comes first; it's the only sane way of doing things based on the experiences this year |
| 20:24 |
celeron55 |
i'm really worried again now that hmmmm isn't around again |
| 20:29 |
VanessaE |
at least he's around. we lost zeno (lack of interest) |
| 20:29 |
celeron55 |
i know; zeno /msg'd me when he left |
| 20:30 |
est31 |
and hmmmm 's at vacation, so he isnt really around either. |
| 20:30 |
celeron55 |
he was quite absent anyway for a long time already so it's not very surprising |
| 20:34 |
celeron55 |
also yes, my farmap schedule is a train wreck |
| 20:35 |
celeron55 |
i just rebased it on upstream/master though so that's somethting |
| 20:35 |
celeron55 |
-t |
| 20:37 |
celeron55 |
my priority is to get it stable before asking it to be reviewed for merging, and frankly the rule of the last 10% being the another 90% applies here |
| 20:37 |
VanessaE |
heh |
| 20:37 |
celeron55 |
i'm not going to even pretend otherwise |
| 20:39 |
celeron55 |
to those arguing about regular mapblock rendering having to be faster: yeah probably; i'm laying some possible groundwork for doing that when doing this so the future of that isn't probably completely hopeless |
| 20:40 |
celeron55 |
GPU-based occlusion culling isn't some kind of magic bullet like some poeple seem to be thinking |
| 20:41 |
celeron55 |
urho3d can do it easily and i tested the performance when developing buildat; it's not a trivial solution in voxel environments |
| 20:41 |
VanessaE |
celeron55: perhaps not, but can it really be argued that a GPU would be slower at that task than regular software? |
| 20:42 |
celeron55 |
you need to take into account the amount of work required for making changes and also the inflexibilities of irrlicht |
| 20:42 |
VanessaE |
(assuming a modern, still-supported-by-OEM GPU of course) |
| 20:43 |
celeron55 |
the amount of skilled work required* |
| 20:45 |
Fixer |
what about per node software oclusion? |
| 20:45 |
est31 |
Fixer, that would be against how stuff works currently |
| 20:45 |
celeron55 |
per node occlusion culling immediately shows that the person doesn't know what they are talking about |
| 20:46 |
rubenwardy |
per node occlusion wouldn't be good, that implies you draw nodes individually |
| 20:46 |
VanessaE |
sure it does. why wipe entires mapblocks of content off the screen if the GPU can render them at a given frame rate? |
| 20:46 |
VanessaE |
-s |
| 20:46 |
Fixer |
for example, if I stand 50 nodes away from spawn and looking at wall in direction of spawn, my fps goes down a lot |
| 20:47 |
VanessaE |
(the converse as stated by fixer applies as well) |
| 20:47 |
VanessaE |
maybe calling it "per node" is the wrong way to think of it then |
| 20:50 |
est31 |
per node occlusion would make sense, yes, but do it via a vertex shader please. |
| 20:50 |
VanessaE |
point is, you've got X number of polys going into the rendering pipeline, and somehow you gotta keep them below Y number of polys for them to render at some given FPS. |
| 20:51 |
celeron55 |
VanessaE: your hardware isn't poly-limited; it's almost purely number-of-drawcalls limited at least when VBOs are used |
| 20:51 |
VanessaE |
whether you divide them into nodes, mapblocks, or chunks or some other measure, Y needs to be as close as possible to X without distorting your view. |
| 20:51 |
celeron55 |
and GPU occlusion culling requires more drawcalls |
| 20:51 |
RealBadAngel |
celeron55, whatever you gonna say about occlusion, forcing client to regen meshes for several mapblocks is just wrong |
| 20:51 |
VanessaE |
celeron55: ok. well I'm just being a bit simplistic here, I'm sure you understand. |
| 20:52 |
RealBadAngel |
with a single meta change... i mean |
| 20:52 |
VanessaE |
of course with MT not using VBO right now... |
| 20:52 |
celeron55 |
RealBadAngel: wrong? well it's wrong if you can fix it properly; not otherwise |
| 20:52 |
VanessaE |
celeron55: he needs help to fix it. |
| 20:53 |
RealBadAngel |
celeron55, its total waste of resources. move somehthing in a chest == regen of 9 mapblocks |
| 20:53 |
RealBadAngel |
its your design? |
| 20:54 |
celeron55 |
the design is rather KISS; it doesn't special-case anything |
| 20:54 |
RealBadAngel |
well that particular KISS KILLS ant complex server... |
| 20:55 |
RealBadAngel |
*any |
| 20:57 |
RealBadAngel |
celeron55, and thats the problem, once it was chosen to send whole mapblock |
| 20:57 |
RealBadAngel |
and it is like that since always |
| 20:57 |
VanessaE |
or more to the point, it is detrimental to any client, particularly slow-CPU ones, who is connecting to a complex server. |
| 20:57 |
RealBadAngel |
now mods plays a lot with meta |
| 20:57 |
RealBadAngel |
entering complex world kills the clients |
| 20:57 |
celeron55 |
RealBadAngel: you don't need to talk about it to me |
| 20:58 |
celeron55 |
RealBadAngel: so you shouldn't |
| 20:58 |
RealBadAngel |
celeron55, so are you aware of how it is important or you just dont care about it? |
| 20:58 |
celeron55 |
RealBadAngel: special casing it probably makes sense; it's out of scope for me |
| 20:59 |
RealBadAngel |
so dont care, right? |
| 20:59 |
celeron55 |
make it, and make it not buggy, and make it compatible and otherwise good too |
| 20:59 |
celeron55 |
don't bug me about it |
| 21:00 |
celeron55 |
i am not your caretaker |
| 21:01 |
VanessaE |
guys, stop.. |
| 21:01 |
RealBadAngel |
ofc not |
| 21:02 |
RealBadAngel |
im just asking here for some help to finish that damn task |
| 21:02 |
RealBadAngel |
if i do realize where the problem is it doesnt mean i can code the solution |
| 21:05 |
celeron55 |
the thing is, if i dig into it enough to help you, i might just as well just fix the whole thing myself |
| 21:06 |
celeron55 |
it's almost the same amount of work, maybe even less |
| 21:06 |
RealBadAngel |
most of the work is already done i think |
| 21:06 |
celeron55 |
i'll rather use my time to focus on what i last started doing and didn't finish yet |
| 21:07 |
RealBadAngel |
ok. i wont push it anymore and will try again to do it on my own |
| 21:07 |
RealBadAngel |
peace? |
| 21:09 |
celeron55 |
this is not a special occasion; this is generally the case when people ask for help; and yes, i don't hold a grudge about this if you understand this |
| 21:09 |
|
sfan5_ joined #minetest-dev |
| 21:10 |
RealBadAngel |
celeron55, ok |
| 21:15 |
|
JohnnyComeL8ly joined #minetest-dev |
| 21:26 |
RealBadAngel |
celeron55, on your farmesh code youre using VBO, have you found a reliable way to make sure that unused meshes got deleted? |
| 21:28 |
RealBadAngel |
thats raher important question, moving mapblock meshes to hw buffers doubles the framerate |
| 21:28 |
|
proller joined #minetest-dev |
| 21:31 |
celeron55 |
at least it never re-creates the meshes in the way mapblock meshes are completely re-created |
| 21:31 |
celeron55 |
but i'm not sure if it has the same problem or not |
| 21:31 |
RealBadAngel |
if they got uploaded theyre stored |
| 21:32 |
RealBadAngel |
so the same problem |
| 21:33 |
RealBadAngel |
youre usin static flag, meaning uploading once |
| 21:34 |
celeron55 |
it would seem that if nothing else works, creating a VBO mesh pool would work, because irrlicht will update them in-place when you mark them dirty (i mean, god help us if it even then creates a duplicate (it won't)) |
| 21:34 |
celeron55 |
there's also IVideoDriver::removeHardwareBuffer |
| 21:34 |
RealBadAngel |
we were playin already with VBO code |
| 21:35 |
RealBadAngel |
flushing the buffers was the problem |
| 21:35 |
celeron55 |
what does "flushing the buffers" even mean? |
| 21:36 |
RealBadAngel |
clear the GPU memory used? idk, VBO idea was killed because of it |
| 21:36 |
|
Supertanker2 joined #minetest-dev |
| 21:37 |
VanessaE |
celeron55: in the old code, over time outdated VBO entries would pile up as mapblocks are changed, and never get cleared out |
| 21:37 |
VanessaE |
effectively creating a memory leak |
| 21:37 |
celeron55 |
yes i know that was the diagnosis back then |
| 21:38 |
celeron55 |
and it will happen now too if you enable VBOs AFAIK |
| 21:38 |
RealBadAngel |
you did |
| 21:38 |
RealBadAngel |
thats why im askin |
| 21:41 |
RealBadAngel |
have you solved memleaks issue with VBO or just put the problem away to be solved later on |
| 21:42 |
celeron55 |
checking it is one of the tens of entries on my far map to-do list |
| 21:42 |
celeron55 |
it hasn't seemed like a huge issue when i have been testing it though |
| 21:42 |
celeron55 |
not sure why, but anyway |
| 21:43 |
RealBadAngel |
that issue was here before your farmesh code |
| 21:43 |
RealBadAngel |
solution to it could be of great surplus |
| 21:44 |
RealBadAngel |
twice fps doesnt walk the street for free ;) |
| 21:45 |
RealBadAngel |
celeron55, ive also havent seen any issues when testing VBO push code (i would like to get and see memleaks about it) |
| 21:47 |
RealBadAngel |
all the sources i read mentioned forced delete with removeHardwareBuffer after some time (not used, timer or whatever) |
| 21:49 |
|
Darcidride joined #minetest-dev |
| 21:49 |
RealBadAngel |
if it is problematic, couldnt it be pushed mainstream for more folks to test it? |
| 21:50 |
RealBadAngel |
we can always revert single change |
| 21:50 |
RealBadAngel |
but feedback is priceless |
| 21:51 |
celeron55 |
if you could link said changeset here, it would be helpful |
| 21:51 |
RealBadAngel |
it was initially PilzAdam's code |
| 21:52 |
celeron55 |
i don't care whose it was or where it originally was; i care where it's available currently |
| 21:52 |
RealBadAngel |
he forgert about setting mesh to dirty in animation |
| 21:52 |
RealBadAngel |
so it was its visible flaw |
| 21:53 |
celeron55 |
many things have changed since then so it might just work for whatever unknown reason now |
| 21:53 |
RealBadAngel |
i can prepare quickly a PR with it |
| 21:53 |
celeron55 |
like, irrlicht 1.8 could have fixed it |
| 21:53 |
celeron55 |
or whatever |
| 21:53 |
RealBadAngel |
thats a few liners |
| 21:54 |
VanessaE |
https://github.com/PilzAdam/minetest/tree/vbo |
| 21:54 |
VanessaE |
last two commits in that branch by the look of it |
| 21:55 |
RealBadAngel |
ah, he really introduced the timer |
| 21:55 |
celeron55 |
if you make a PR, label it as "request for testing, not for merging"; that should give us some information we need |
| 21:55 |
RealBadAngel |
from what ive read thats good |
| 21:57 |
RealBadAngel |
in this case request for testing should be screamed out loud |
| 21:57 |
RealBadAngel |
on every platform |
| 21:57 |
Obani |
Guys, I've got a problem |
| 21:57 |
celeron55 |
a windows build could be made possibly |
| 21:57 |
RealBadAngel |
oh, NO |
| 21:57 |
Obani |
I compiled Minetest with latest commit |
| 21:57 |
celeron55 |
i'm sure people will test it if you suggest it might give more FPS 8) |
| 21:57 |
Obani |
An got this error at launching http://pastebin.com/M9rQhpqi |
| 21:57 |
Obani |
(I cannot launch the game because of a segmentation error) |
| 21:58 |
RealBadAngel |
celeron55, it doubles the fps, believe me |
| 21:58 |
celeron55 |
RealBadAngel: tell it to testers |
| 21:58 |
RealBadAngel |
oke doke, i will |
| 21:59 |
Fixer |
i will test it, just give me windows binary |
| 21:59 |
nrzkt |
Fix should be backported to current minetest |
| 22:00 |
RealBadAngel |
not yet, hold on please. maybe tommorow |
| 22:00 |
nrzkt |
using the branch as is is not a test :) |
| 22:00 |
JohnnyComeL8ly |
Fixer, use Linux... that is the best way to test. |
| 22:00 |
RealBadAngel |
celeron55, i do have another question: https://github.com/Jeija/minetest/tree/clientside_mods |
| 22:01 |
RealBadAngel |
why that was not in? |
| 22:01 |
RealBadAngel |
asking because i do need that funcionality |
| 22:01 |
RealBadAngel |
at least on startup |
| 22:02 |
RealBadAngel |
you have mentioned lately client side mods |
| 22:03 |
Obani |
"Audio: Vorbis extension NOT present" |
| 22:03 |
Obani |
What does it mean ? |
| 22:04 |
Obani |
A missing package ? |
| 22:04 |
Obani |
lib ? |
| 22:04 |
JohnnyComeL8ly |
Maybe you didn't compile with Vorbis support.... |
| 22:04 |
JohnnyComeL8ly |
Yes, something like that. |
| 22:04 |
Obani |
ok... |
| 22:04 |
JohnnyComeL8ly |
Didn't you install all the deps? |
| 22:05 |
Obani |
??? |
| 22:05 |
Obani |
It crashes becaus ether is no main_menu music ? 0.o |
| 22:05 |
JohnnyComeL8ly |
No. |
| 22:06 |
JohnnyComeL8ly |
What did you install? |
| 22:06 |
RealBadAngel |
main menu file can show you just a warning |
| 22:06 |
RealBadAngel |
cannot be a reason to crash |
| 22:06 |
JohnnyComeL8ly |
The exact same list as in https://github.com/minetest/minetest/blob/master/README.txt |
| 22:07 |
Obani |
Look, it as been working for 3 years |
| 22:07 |
Obani |
Why should it crash now ? xD |
| 22:09 |
JohnnyComeL8ly |
Obani... did you install all the deps? |
| 22:09 |
Obani |
I'm doing it right now |
| 22:09 |
Obani |
Some were added :p |
| 22:09 |
Obani |
(since last time) |
| 22:09 |
JohnnyComeL8ly |
Well, that could be our problem. |
| 22:11 |
Obani |
Let's recompile it... yay ! |
| 22:12 |
JohnnyComeL8ly |
I'm excited too. |
| 22:15 |
RealBadAngel |
celeron55, whats your word on jeija's PR ? |
| 22:17 |
Obani |
Still got the same errors |
| 22:17 |
Obani |
After installing the depencies and recompiling |
| 22:18 |
JohnnyComeL8ly |
Did you do a complete rebuild or were there things that didn't get rebuilt? |
| 22:19 |
JohnnyComeL8ly |
Obani ^^ |
| 22:19 |
Obani |
I did a complet rebuilt |
| 22:19 |
Obani |
complete* |
| 22:19 |
Obani |
JohnnyComeL8ly, yes ? |
| 22:20 |
JohnnyComeL8ly |
I'm not able to explain it properly, but there are things that are built before it is all tied together, and some of those things might be "old" |
| 22:21 |
Obani |
JohnnyComeL8ly, I should reclone the git repository ? |
| 22:22 |
|
rubenwardy joined #minetest-dev |
| 22:29 |
JohnnyComeL8ly |
Obani, I'm thinking more about the git clean -xf cmd |
| 22:29 |
Obani |
too late |
| 22:29 |
Obani |
already compiling |
| 22:29 |
JohnnyComeL8ly |
Oh, well. |
| 22:44 |
Obani |
Still doesn't work |
| 22:46 |
Obani |
(And still same errors) |
| 22:48 |
JohnnyComeL8ly |
Try getting a completely fresh clone... I don't know for sure how to get a clean source. |
| 22:48 |
RealBadAngel |
make new folder |
| 22:48 |
JohnnyComeL8ly |
(So that is what gives 'til somebody helps) |
| 22:48 |
RealBadAngel |
and grab the sources |
| 22:49 |
RealBadAngel |
unzip them there |
| 22:49 |
RealBadAngel |
and make sure youre compiling with -DRUN_IN_PLACE=1 |
| 22:50 |
RealBadAngel |
this ways each folder you will get working mt |
| 22:51 |
Obani |
RealBadAngel, still doesn't work |
| 22:51 |
RealBadAngel |
paste bin output you get |
| 22:51 |
Obani |
Already did it |
| 22:51 |
Obani |
It was this http://pastebin.com/M9rQhpqi |
| 22:51 |
Obani |
(I used --trace |
| 22:53 |
RealBadAngel |
skip trace |
| 22:54 |
Obani |
lol |
| 22:54 |
Obani |
Ok xD |
| 22:54 |
RealBadAngel |
idk just a guess |
| 22:54 |
Obani |
Loaded texture: /home/obani/minetest/games/minetest_game/menu/header.png |
| 22:54 |
Obani |
:D |
| 22:56 |
Obani |
(no offense to you, I just tell that I have those same errors since some hours) |
| 22:58 |
JohnnyComeL8ly |
I'm gonna switch to Xubuntu and try a fresh compile. |
| 22:59 |
RealBadAngel |
Obani, i cant see any errors just warnings |
| 23:00 |
Obani |
Yes, I noticed that |
| 23:00 |
VanessaE |
current HEAD builds and runs fine for me on xubuntu 15.10. |
| 23:00 |
VanessaE |
sound too. |
| 23:01 |
RealBadAngel |
maybe deps then |
| 23:01 |
Obani |
I'm not on xubuntu |
| 23:01 |
Obani |
NO IT'S NOT DEPS |
| 23:01 |
VanessaE |
it still may be, if some lib is too old or too new for MT to use. |
| 23:02 |
Obani |
sorry night affects me |
| 23:02 |
Obani |
VanessaE, So MT is not abled to work on Debian stable |
| 23:02 |
RealBadAngel |
are you a vampire or what? :P |
| 23:02 |
VanessaE |
Obani: I could easily go check that... |
| 23:03 |
VanessaE |
clone... |
| 23:03 |
RealBadAngel |
github check will fail then.... |
| 23:04 |
VanessaE |
building... |
| 23:04 |
RealBadAngel |
obani can you pastebin output of cmake . -DRUN_IN_PLACE=1 ? |
| 23:05 |
Obani |
RealBadAngel, do I have to recompile it again ? |
| 23:05 |
VanessaE |
no probs compiling. |
| 23:05 |
RealBadAngel |
delete cmakecache.txt |
| 23:05 |
RealBadAngel |
and call cmake again |
| 23:05 |
VanessaE |
(not gonna run it though, too lazy to set up a proper test environ) |
| 23:06 |
Obani |
http://pastebin.com/SjG0daMP |
| 23:06 |
Obani |
Ah |
| 23:06 |
Obani |
ok wait |
| 23:07 |
RealBadAngel |
VanessaE, btw have you saw mt logo as mesh? |
| 23:07 |
VanessaE |
no |
| 23:07 |
Obani |
RealBadAngel, here it is http://pastebin.com/1szuA5vi |
| 23:09 |
RealBadAngel |
Obani, now compile it again and log please |
| 23:09 |
VanessaE |
lgtm |
| 23:09 |
RealBadAngel |
VanessaE, it was posted on the forums lately |
| 23:09 |
VanessaE |
RealBadAngel: link? |
| 23:10 |
Obani |
VanessaE, WIP mods |
| 23:11 |
RealBadAngel |
im still lookin, some weird place it was |
| 23:12 |
RealBadAngel |
https://forum.minetest.net/viewtopic.php?f=9&t=7034 |
| 23:12 |
RealBadAngel |
yeah |
| 23:13 |
RealBadAngel |
the problem is the model is textured somehow weird |
| 23:13 |
VanessaE |
ah |
| 23:13 |
RealBadAngel |
i need the mesh as an obj and with texture |
| 23:14 |
RealBadAngel |
i want it to spin in main menu, in the background |
| 23:14 |
Obani |
This model ? |
| 23:14 |
Obani |
That would look a bit weird IMO |
| 23:15 |
RealBadAngel |
and yes, i do have code for meshes in the formspecs |
| 23:15 |
VanessaE |
RealBadAngel: pulling from the textures the current texture pack uses, I hope ;) |
| 23:15 |
* VanessaE |
imagines an HDX version of that logo |
| 23:16 |
RealBadAngel |
main menu should be something bigger than hard to point set of burried options |
| 23:17 |
RealBadAngel |
what we do have atm is nothing like a main should look like |
| 23:17 |
Obani |
I agree with that, but a single mesh turning in the background would look a bit special ^^ |
| 23:18 |
RealBadAngel |
Obani, im a demoscene coder |
| 23:18 |
RealBadAngel |
leave that to me, please |
| 23:18 |
Obani |
BUT |
| 23:19 |
VanessaE |
no buts :) |
| 23:19 |
Obani |
a 3d land like MT mesh with animated things on it looping would be cool ;) |
| 23:19 |
Obani |
But much more difficult to make :p |
| 23:19 |
RealBadAngel |
i havent said that mesh will be everthin |
| 23:19 |
RealBadAngel |
i said i just need it |
| 23:20 |
Obani |
Also, what are the plans for graphical enhancements on MT ? |
| 23:20 |
RealBadAngel |
client side mods |
| 23:21 |
Obani |
"graphical" |
| 23:21 |
RealBadAngel |
before that you can try server side ones |
| 23:21 |
RealBadAngel |
like water or lava |
| 23:21 |
Obani |
client-side mods : mods are requested by server and then used by client ? |
| 23:25 |
Obani |
RealBadAngel, still the same http://pastebin.com/3XkKtk3s |
| 23:26 |
RealBadAngel |
Obani, client side mods means, chosen by client and server indenpenent |
| 23:27 |
RealBadAngel |
like texture packs |
| 23:27 |
VanessaE |
RealBadAngel: it also means server sends mod code to be run on the client. |
| 23:27 |
Obani |
RealBadAngel, so that would enable cheat ? |
| 23:27 |
|
Wayward_One joined #minetest-dev |
| 23:28 |
RealBadAngel |
for starters i think once executed mods, on startup should be enough |
| 23:28 |
RealBadAngel |
just to modify nodedefs |
| 23:29 |
RealBadAngel |
security is another problem and i dont have any opinion on that atm, not my focus |
| 23:29 |
|
TeTpaAka joined #minetest-dev |
| 23:30 |
RealBadAngel |
jeija's proof of concept allows dtime run mods |
| 23:31 |
Obani |
Well, add a server security in MT.conf would be the first step |
| 23:31 |
RealBadAngel |
i really wonder why that code went unnoticed for 2 yrs |
| 23:31 |
Obani |
enable_mods = default wool fire ... ... |
| 23:32 |
RealBadAngel |
Obani, i wont talk about securtity, i do care only what i can alter client side regarding rendering |
| 23:32 |
RealBadAngel |
i do need that code for client side shaders |
| 23:33 |
RealBadAngel |
what ive coded water and lava shaders propably wont go in into main repo |
| 23:33 |
nrzkt |
VanessaE i think first step is only customizing things like gui, it's better to have a separated world first time before having an API permitting to hack every client without restriction :p |
| 23:34 |
RealBadAngel |
its much of an artistic vision than a voxel one, so i need to make a way clients use it on their choice |
| 23:34 |
RealBadAngel |
client side modding is just a proper way to achieve that goal |
| 23:35 |
RealBadAngel |
and not only me will use it |
| 23:35 |
TeTpaAka |
Is there anyone around who could help me with https://github.com/minetest/minetest/pull/2738 ? |
| 23:36 |
RealBadAngel |
make Taoki test it |
| 23:37 |
RealBadAngel |
hes the one to wash it out |
| 23:37 |
TeTpaAka |
I don't know if he tested it. I already showed it to him and he commented on it. |
| 23:38 |
RealBadAngel |
commented that he would love it to have |
| 23:38 |
RealBadAngel |
make him test it actually |
| 23:38 |
RealBadAngel |
Creatures mod will benefit that change a lot |
| 23:39 |
RealBadAngel |
also players crouching too |
| 23:39 |
|
CraigyDavi left #minetest-dev |
| 23:39 |
TeTpaAka |
That would be out of scope for the pull |
| 23:39 |
RealBadAngel |
but will allow that |
| 23:39 |
TeTpaAka |
And the pull works, that is not the problem |
| 23:39 |
TeTpaAka |
The problem is the hackiness of the code |
| 23:40 |
TeTpaAka |
It enlarges a problem of the positioning of the player |
| 23:40 |
TeTpaAka |
The player model has an offset to the mob models |
| 23:40 |
TeTpaAka |
It doesn't stand on (0, 0, 0) has that point in the chest area. |
| 23:40 |
RealBadAngel |
if it works as intended thats not a hack |
| 23:40 |
RealBadAngel |
just a black magic ;) |
| 23:41 |
TeTpaAka |
The problem is, that for the mod programmer it is completely opaque, why he needs to offset the player model when he uses the same model as for mods |
| 23:42 |
TeTpaAka |
Ideally he would use the same models |
| 23:42 |
TeTpaAka |
But this would break every existing mod that relies on these features. |
| 23:42 |
RealBadAngel |
docuemnt the differences |
| 23:42 |
RealBadAngel |
and mark whats needed as TODO |
| 23:42 |
TeTpaAka |
I didn't introduce any differences. These differences already exist. |
| 23:43 |
RealBadAngel |
so mark them |
| 23:58 |
|
behalebabo joined #minetest-dev |
| 23:59 |
|
proller joined #minetest-dev |
