| Time |
Nick |
Message |
| 00:03 |
RealBadAngel |
well, you were seriously adding noise to textures> |
| 00:03 |
RealBadAngel |
? |
| 00:03 |
Taoki |
You mean my idea about noise on solid block vertices? |
| 00:04 |
RealBadAngel |
yes |
| 00:04 |
Taoki |
Well I didn't mean noise to textures. This would probably have nothing to do with textures really |
| 00:04 |
Taoki |
I meant noise to vertices |
| 00:04 |
Taoki |
So the ground can look rough a bit, and more realistic for who wants that |
| 00:05 |
Taoki |
Similar to how the water shader creates waves, but of course static and with more detail |
| 00:05 |
RealBadAngel |
well, that parallax mapping does already |
| 00:06 |
Taoki |
Can only do it to a small extent yeah |
| 00:06 |
Taoki |
Was thinking of this as an additional effect, for the actual surface to be bent and rough |
| 00:06 |
Taoki |
Like this http://minecraftpremiumfree.net/premium/minecraft_logo.png |
| 00:06 |
RealBadAngel |
have you saw any implementation of similar idea? some pics? |
| 00:07 |
Taoki |
Or better, like this: http://collectionminecraft.com/wp-content/uploads/2014/03/minecraft-grass-block-iconimage-from-2--minecraft-block-post-3-kquoyplt.png Excluding grass particles on top, but see the edges |
| 00:08 |
Taoki |
I only know it's possible for shaders to deform vertices. Possibly subdivide them a bit |
| 00:08 |
Taoki |
Water already does that as it wavers now |
| 00:09 |
|
grrk-bzzt_ joined #minetest-dev |
| 00:10 |
Taoki |
http://p1.pichost.me/i/27/1497704.jpg If only something like this were possible in-game :) |
| 00:11 |
Taoki |
^ That's another idea of block smoothing I have in mind. See the round edges |
| 00:11 |
Taoki |
Rounding the edges of blocks a bit felt like another nice dea ^^ But prolly harder to do... still felt like talking about it |
| 00:13 |
RealBadAngel |
http://www.deepmesh3d.com/help/matblend.htm |
| 00:13 |
RealBadAngel |
take a look on possible material types |
| 00:14 |
Taoki |
http://gameslikeminecraft.ddl-it.com/wp-content/uploads/2012/08/grass.png Ahhh... if only... |
| 00:14 |
Taoki |
ok |
| 00:18 |
RealBadAngel |
not possible |
| 00:19 |
RealBadAngel |
that is made out of polygons, not flat textures |
| 00:19 |
Taoki |
RealBadAngel: Yeah, those are some very nice material possibilities |
| 00:19 |
* Taoki |
nods |
| 00:27 |
Taoki |
Guess I like to think of more than is possible sometimes. don't automatically take what I say as a feature request or anything though... I'm not that crazy :P Though roughing the vertices of solid objects was an actual idea... clearly nothing urgent or important now tho |
| 00:39 |
Megaf |
RealBadAngel: the reddish thing happens when I enable waving something |
| 00:39 |
Megaf |
and it happens only with your branch |
| 00:40 |
Megaf |
on normal minetest master branch I have everything waving and all shaders |
| 00:43 |
Megaf |
RealBadAngel: just confirmed here, all shaders do work fine in minetest master branch |
| 00:50 |
RealBadAngel |
Megaf, ive located the bug, thx |
| 00:51 |
RealBadAngel |
i will push the changes shotly |
| 00:51 |
RealBadAngel |
+r |
| 00:52 |
Taoki |
RealBadAngel: Isn't a gradient map included in minetest_game by default? Was curious to change the colors and test it a bit |
| 00:52 |
Taoki |
To see how and if it all works |
| 00:52 |
Taoki |
I mean, for coloring the fog and all |
| 00:53 |
RealBadAngel |
hmmm, tonemaps are used in mc |
| 00:53 |
RealBadAngel |
for grass and lotsa other things |
| 00:54 |
RealBadAngel |
if you mean fog i suggest using the one i made for sun |
| 00:54 |
RealBadAngel |
this way fog and sun could have very same colors |
| 00:54 |
Taoki |
RealBadAngel: Ok. Where can I get the image file for it? And where do I place it? Does it work automatically also, without a Lua function? |
| 00:55 |
RealBadAngel |
hold on a sec |
| 00:55 |
RealBadAngel |
https://forum.minetest.net/viewtopic.php?f=18&t=9140 |
| 00:55 |
RealBadAngel |
grab tonemaps from here |
| 00:56 |
Taoki |
Thanks |
| 00:56 |
RealBadAngel |
but, for tonemaps to work, you have to use textures in the first place |
| 00:56 |
RealBadAngel |
and its automatic |
| 00:56 |
|
daswort joined #minetest-dev |
| 00:56 |
Taoki |
If I can get the sun and moon tonemap as a RGB color in sky.cpp, fog colors are as good as solved |
| 00:57 |
RealBadAngel |
just put those files either in textures pack folder or /textures/default/pack |
| 00:57 |
Taoki |
Also, texturable sun and moon mean that you can texture the sun and moon themselves, rather than them being blocks, right? |
| 00:57 |
Taoki |
So they can be images |
| 00:57 |
Taoki |
Are they still tonemap colored then also? |
| 00:58 |
RealBadAngel |
if you will give sun a texture it will use it |
| 00:58 |
RealBadAngel |
if you will put there tonemap it will be tinted also |
| 00:58 |
Taoki |
Nice |
| 00:58 |
Taoki |
Yeah was just curious if sun and moon can use texture images now. Good to know |
| 00:59 |
RealBadAngel |
hold on |
| 00:59 |
RealBadAngel |
https://www.youtube.com/watch?v=hubBBzh_190 |
| 00:59 |
Taoki |
Ah, cool... yes it works then |
| 01:05 |
RealBadAngel |
yes |
| 01:05 |
RealBadAngel |
and i think tonemap is quite good |
| 01:05 |
RealBadAngel |
it blends nicely with dawn/dusk image |
| 01:08 |
Taoki |
How are the tonemap colors named in the code? So I can know what to get in sky.cpp |
| 01:08 |
RealBadAngel |
color is picked from file using some blac magic |
| 01:09 |
Taoki |
At startup I assume, but not dynamically right? |
| 01:09 |
RealBadAngel |
https://github.com/minetest/minetest/blob/master/src/sky.cpp#L166 |
| 01:10 |
RealBadAngel |
dynamically |
| 01:10 |
RealBadAngel |
depending on time of day |
| 01:10 |
RealBadAngel |
oops, https://github.com/minetest/minetest/blob/master/src/sky.cpp#L164 |
| 01:11 |
RealBadAngel |
at 164 offset is calculated |
| 01:11 |
RealBadAngel |
sin is used there to make the changes faster at start |
| 01:11 |
RealBadAngel |
then slower |
| 01:12 |
Taoki |
m_materials[3].EmissiveColor and m_materials[4].EmissiveColor seem to be the two colors. Will see |
| 01:13 |
Taoki |
Need to go now, will look at it tomorrow |
| 01:13 |
Taoki |
night |
| 01:13 |
RealBadAngel |
3 is sun, 4 is moon |
| 01:13 |
RealBadAngel |
cya |
| 01:14 |
Megaf |
RealBadAngel: wheres my fix? :) |
| 01:15 |
RealBadAngel |
about to test it, gimme 15 minutes |
| 01:21 |
RealBadAngel |
Megaf, pushed |
| 01:22 |
RealBadAngel |
but you will need one extra file for that |
| 01:22 |
RealBadAngel |
lemme upload it somwhere |
| 01:26 |
|
kaeza joined #minetest-dev |
| 01:27 |
RealBadAngel |
http://i.imgur.com/iW5TY2F.png |
| 01:27 |
RealBadAngel |
save it as water_surface_animated_normal.png in any textures path |
| 01:28 |
RealBadAngel |
this file is used by water surface shader |
| 01:31 |
ShadowNinja |
hmmmm: hexafraction was saying that he planned to do exactly that (two nodes per MapBlock, one with translucent nodes) in mosstest (a Minetest clone in Java). |
| 01:33 |
hmmmm |
cool then, but I'd rather wait for celeron55's comment on it |
| 01:33 |
hmmmm |
if it were viable I'm sure he would've done that |
| 01:38 |
RealBadAngel |
since ive done shader separation and move them all to tiles |
| 01:39 |
RealBadAngel |
thats no problem for me to introduce more scene nodes |
| 01:39 |
RealBadAngel |
i couldnt do that before because everything were bound to tightly |
| 01:42 |
RealBadAngel |
branch name "united" can be misleading in fact |
| 01:43 |
RealBadAngel |
theyre unitied because theres one shadar for all the nodes |
| 01:43 |
RealBadAngel |
but it is preconfigured for all the possible materials and drawtypes |
| 01:43 |
RealBadAngel |
at the compile time |
| 01:44 |
RealBadAngel |
also my latest push is about to add totally different shader, water surface one |
| 01:46 |
RealBadAngel |
also i do have (almost) rdy lava surface shader that will use bezier curves to create lava rivers/flowing lava surfaces |
| 01:46 |
|
daswort joined #minetest-dev |
| 01:47 |
VanessaE |
hmmmm: celeron's comment on what, doing one node per mapblock? |
| 01:47 |
hmmmm |
yes |
| 01:47 |
RealBadAngel |
that wont be any good |
| 01:47 |
VanessaE |
[04-29 19:12] <celeron55> 01:53:48 <+hmmmm> what problems are there with making a mapblock an entire node? |
| 01:47 |
VanessaE |
[04-29 19:12] <celeron55> none, assuming you have infinite time for implementing it |
| 01:48 |
hmmmm |
Oh |
| 01:48 |
hmmmm |
i didn't see that |
| 01:48 |
RealBadAngel |
what we do need is one node per type |
| 01:48 |
RealBadAngel |
one water, one dirt etc |
| 01:49 |
RealBadAngel |
anything else is just clusterfuck, same as we do already have |
| 01:49 |
hmmmm |
that might fix transparency but that'd defeat the purpose |
| 01:49 |
hmmmm |
the main purpose of doing this is to add occlusion culling and query |
| 01:49 |
RealBadAngel |
^^ above |
| 01:50 |
hmmmm |
huh |
| 01:50 |
hmmmm |
anyway I don't get what celeron means by that |
| 01:50 |
hmmmm |
is it very difficult to implement or something? |
| 01:50 |
RealBadAngel |
not really |
| 01:51 |
VanessaE |
[04-29 19:12] <celeron55> or, well |
| 01:51 |
VanessaE |
[04-29 19:13] <celeron55> you have to somehow build into it the occlusion culling and so |
| 01:51 |
VanessaE |
[04-29 19:13] <celeron55> but it shouldn't be a big deal |
| 01:51 |
VanessaE |
so idk |
| 01:51 |
RealBadAngel |
in fact i could code in a few days |
| 01:52 |
VanessaE |
I got the impression that it would be incredibly non-trivial to implement |
| 01:52 |
RealBadAngel |
last few months for me were rather reading irrlicht docs and sample apps |
| 01:52 |
VanessaE |
but it begs the question: how fine-grained does the occlusion culling need to get to get the most benefit? |
| 01:53 |
RealBadAngel |
VanessaE, in fact u shouldnt even bother about it |
| 01:53 |
RealBadAngel |
irrlicht does everything for u |
| 01:56 |
VanessaE |
perhaps. I don't understand irrlicht enough to know just how much work it would have to do with a mapblock-sized irrlicht node versus, say, one node per material per mapblock |
| 01:56 |
|
[PavelS] joined #minetest-dev |
| 02:00 |
hmmmm |
hmm |
| 02:00 |
|
OldCoder joined #minetest-dev |
| 02:00 |
hmmmm |
i wonder if there is a disadvantage to having too many scene nodes |
| 02:01 |
VanessaE |
dunno |
| 02:01 |
VanessaE |
the way I understood it when i read up on this before, a lot is better than few, but I didn't see anything that said one way or the other about any practical limits |
| 02:04 |
VanessaE |
one per mapblock seems perfectly reasonable |
| 02:05 |
VanessaE |
(well one per transparent/translucent/opaque/liquid, let's say) |
| 02:05 |
hmmmm |
yeah |
| 02:05 |
hmmmm |
anyway I'll look at it when i get the time |
| 02:06 |
|
daswort joined #minetest-dev |
| 02:07 |
VanessaE |
hm, with a view range of 100, 4/3 pi R ^3 makes a maximum volume of around 4188790 nodes or about 1023 mapblocks |
| 02:07 |
VanessaE |
(if every last mapblock had non-air) |
| 02:08 |
Megaf |
RealBadAngel: that red thing is fixed indeed, I still get 4 lines of this 23:07:14: ERROR[main]: Reading from file! |
| 02:08 |
VanessaE |
Megaf: ignore that. |
| 02:08 |
VanessaE |
it's not a bug, it's a diagnostic feature. |
| 02:08 |
Megaf |
ok |
| 02:09 |
Megaf |
and water is white |
| 02:09 |
Megaf |
feature too? |
| 02:09 |
VanessaE |
just a diagnostic. ignore it. |
| 02:11 |
Megaf |
well done RealBadAngel. Lights look pretty damn good now |
| 02:11 |
VanessaE |
hmmmm: I don't know how to calculate the visible area from the player's FoV but I guess 20% of that figure maybe? So even one irrlicht node per mapblock would probably barely even touch even the slowest CPU. If there's a benefit to doing one per material type. seems like there's tons of room to do that. |
| 02:11 |
Megaf |
everything is beautiful |
| 02:12 |
Megaf |
everythin is on, 10 FPS at 1280x720 |
| 02:12 |
VanessaE |
RealBadAngel: btw, check lava lighting. sources are lit differently from flowing again (in master) |
| 02:21 |
VanessaE |
um, guys... |
| 02:21 |
VanessaE |
Megaf: you know that issue you had with timeout messages leading to server lockouts? |
| 02:21 |
VanessaE |
it's suddenly happening here too on one of my servers |
| 02:22 |
* VanessaE |
updates... |
| 02:29 |
VanessaE |
For reference, I mean errors of this nature: |
| 02:29 |
VanessaE |
22:20:57: ERROR[ConnectionSend]: con(5/1)RunTimeouts(): Peer 363 has timed out. (source=peer->timeout_counter) |
| 02:30 |
VanessaE |
updated and restarted. It responded to a SIGINT but didn't actually die after about 3 minutes waiting, so I SIGTERM'd it. |
| 02:33 |
Megaf |
VanessaE: yep, some say they are caused by tablets |
| 02:34 |
VanessaE |
strange, this is the first time I've seen it happen though |
| 02:35 |
VanessaE |
well it's back now, and updated. if it happened again I'll file an issue. |
| 02:35 |
VanessaE |
happens* |
| 02:35 |
hmmmm |
VanessaE: I suppose you can model the entire viewable area as a sphere, and the view fulcrum as a rectangle projected onto that sphere |
| 02:36 |
VanessaE |
hmmmm: yeah, that's kinda what I was going for, just to keep the math simple :P |
| 02:36 |
hmmmm |
if you know both the vertical and horizontal FoV you can figure it out |
| 02:36 |
Megaf |
RealBadAngel: when I enable waving water, water waves, but it turns translucid white, and when I disable waving water it turns red |
| 02:36 |
|
daswort joined #minetest-dev |
| 02:36 |
VanessaE |
I know the horizontal is like 90 or 100 degrees |
| 02:36 |
VanessaE |
but no clue about the vertical |
| 02:37 |
VanessaE |
I just don't know the formula. no matter, it's not important anyway |
| 02:44 |
Megaf |
good night |
| 03:02 |
|
daswort joined #minetest-dev |
| 03:02 |
|
werwerwer joined #minetest-dev |
| 03:22 |
|
CraigyDavi joined #minetest-dev |
| 03:30 |
|
daswort joined #minetest-dev |
| 03:53 |
|
WindHero joined #minetest-dev |
| 03:55 |
|
WindHero left #minetest-dev |
| 03:56 |
|
lemon joined #minetest-dev |
| 03:57 |
|
RealBadAngel joined #minetest-dev |
| 04:40 |
|
daswort joined #minetest-dev |
| 04:40 |
|
daswort joined #minetest-dev |
| 04:56 |
|
Miner_48er joined #minetest-dev |
| 06:02 |
|
nore joined #minetest-dev |
| 07:01 |
|
darkrose joined #minetest-dev |
| 07:23 |
|
rsiska joined #minetest-dev |
| 08:38 |
|
ImQ009 joined #minetest-dev |
| 08:49 |
|
OldCoder joined #minetest-dev |
| 09:25 |
|
rsiska joined #minetest-dev |
| 09:27 |
|
PenguinDad joined #minetest-dev |
| 09:43 |
|
RealBadAngel joined #minetest-dev |
| 10:21 |
|
PilzAdam joined #minetest-dev |
| 10:55 |
|
proller joined #minetest-dev |
| 11:14 |
|
Shardvexz joined #minetest-dev |
| 11:18 |
|
troller joined #minetest-dev |
| 11:27 |
|
PenguinDad joined #minetest-dev |
| 11:31 |
|
proller__ joined #minetest-dev |
| 11:34 |
|
Shardvexz joined #minetest-dev |
| 12:08 |
|
grrk-bzzt joined #minetest-dev |
| 12:34 |
Megaf |
folks, we need to implement radio streaming on client side, server would send to client the stream url and then client will get stream from there |
| 12:34 |
Megaf |
this way we could have radio on our servers!! |
| 12:34 |
Megaf |
isnt that cool? |
| 12:37 |
proller__ |
make pull request |
| 12:37 |
Taoki |
RealBadAngel: I'm trying out the tonemaps, but they do not work. I put the textures in the right spot, but it's still the old behavior entirely. Textured sun and moon don't appear to work either |
| 12:38 |
Taoki |
Tried putting the texture in various locations, same result |
| 12:38 |
RealBadAngel |
new texture pack folder should do |
| 12:39 |
RealBadAngel |
how have u named files? |
| 12:39 |
PenguinDad |
Taoki: you need sun and moon textures in order to make tonemaps work |
| 12:40 |
Taoki |
Yes. games/minetest_game/mods/default/textures/moon.png, games/minetest_game/mods/default/textures/moon_tonemap.png, games/minetest_game/mods/default/textures/sun.png, games/minetest_game/mods/default/textures/sun_tonemap.png |
| 12:40 |
Taoki |
Tried in git/textures also, but still nothing |
| 12:40 |
Taoki |
Yeah I got the sun / moon textures too |
| 12:41 |
PenguinDad |
put them in textures/base/pack/ |
| 12:41 |
Taoki |
Directly, or textures/base/pack/default/ ? |
| 12:42 |
PenguinDad |
directly |
| 12:42 |
Taoki |
Ok, now they do work at last. I understood I had to put them in default |
| 12:44 |
Megaf |
proller__: is it possible to get some kind of streaming using a lua mod? |
| 12:46 |
|
[PavelS] joined #minetest-dev |
| 12:48 |
RealBadAngel |
taoki, looks like you have missed /pack/ part.... |
| 12:48 |
Taoki |
right |
| 12:49 |
RealBadAngel |
if you want to test tonemaps be aware that they work if the textures are present |
| 12:50 |
RealBadAngel |
and for them textures shall be prepared slightly different way |
| 12:50 |
RealBadAngel |
the best results you can get with all white textures |
| 12:52 |
RealBadAngel |
megaf, im not sure but that should work without any additions |
| 12:52 |
RealBadAngel |
just make a mod that plays an ogg |
| 12:52 |
Megaf |
RealBadAngel: well, minetest supports ogg |
| 12:53 |
Megaf |
so perhaps an ogg radio? |
| 12:53 |
RealBadAngel |
and use streaming to ogg |
| 12:53 |
RealBadAngel |
there are apps for that |
| 12:53 |
RealBadAngel |
just make sure youre streaming to the file engine will use |
| 12:54 |
RealBadAngel |
openal should do the trick |
| 12:55 |
RealBadAngel |
a few months ago i made minetest playing some net radio broadcast |
| 12:55 |
RealBadAngel |
so it works |
| 12:56 |
PenguinDad |
but every client would have to download the stream :/ |
| 12:56 |
RealBadAngel |
ofc |
| 12:56 |
RealBadAngel |
thats client side |
| 12:57 |
RealBadAngel |
one could even implement there youtube |
| 12:57 |
RealBadAngel |
that could be fun |
| 12:57 |
RealBadAngel |
hey |
| 12:57 |
RealBadAngel |
lets make web browser :) |
| 12:58 |
Anchakor |
you can use browser for GUI :) |
| 12:58 |
RealBadAngel |
i mean real one |
| 12:58 |
|
Fresh_me_ joined #minetest-dev |
| 12:58 |
Anchakor |
I mean real one too |
| 12:58 |
PenguinDad |
let's make tessertest :D |
| 12:58 |
RealBadAngel |
so lets do it |
| 12:59 |
RealBadAngel |
that will be real advantage over mc |
| 12:59 |
Anchakor |
http://www.awesomium.com/ |
| 12:59 |
Anchakor |
free open source alternative is https://code.google.com/p/chromiumembedded/ |
| 13:00 |
Anchakor |
(which is really what awesomium is a wrapper of) |
| 13:00 |
RealBadAngel |
i think i will have to code that terminal thingy |
| 13:00 |
RealBadAngel |
i mean formspec element that will emulate gfx card |
| 13:01 |
RealBadAngel |
with own memory and basic methods to draw/print/whatever |
| 13:02 |
RealBadAngel |
so i can run there CP/M machine |
| 13:06 |
Taoki |
How do you convert from SColor to Scolorf? I only see examples of doing it the other way around |
| 13:10 |
|
hmmmm joined #minetest-dev |
| 13:11 |
|
troller joined #minetest-dev |
| 13:20 |
RealBadAngel |
why would you need to convert that? |
| 13:21 |
RealBadAngel |
please do use f when u can |
| 13:21 |
RealBadAngel |
since colors are floats anyway |
| 13:23 |
Taoki |
RealBadAngel: SColor ranges from 1 to 255 while SColorf from 0 to 1 right? |
| 13:24 |
Taoki |
Getting a strange little bug with colors not being well which is why I was wondering |
| 13:28 |
|
rubenwardy joined #minetest-dev |
| 13:44 |
|
rsiska joined #minetest-dev |
| 13:47 |
|
troller joined #minetest-dev |
| 13:50 |
Taoki |
RealBadAngel: https://github.com/minetest/minetest/pull/1266 All yours to merge |
| 13:52 |
Taoki |
http://i59.tinypic.com/2mn22hs.png |
| 13:52 |
Taoki |
http://i59.tinypic.com/262wtwn.png |
| 13:52 |
PenguinDad |
~title |
| 13:52 |
ShadowBot |
PenguinDad: That URL appears to have no HTML title. |
| 13:52 |
PenguinDad |
~title https://github.com/minetest/minetest/pull/1266 |
| 13:52 |
ShadowBot |
PenguinDad: Make directional fog colors respect tonemap by MirceaKitsune · Pull Request #1266 · minetest/minetest · GitHub |
| 14:00 |
|
proller joined #minetest-dev |
| 14:16 |
|
Shardvexz joined #minetest-dev |
| 14:18 |
Taoki |
RealBadAngel: Everything good with that on your end? |
| 14:19 |
Anchakor |
why people use such terrible ad-ridden image hosting sites like tinypic? |
| 14:20 |
VanessaE |
ads? |
| 14:20 |
VanessaE |
use an adblocker. |
| 14:21 |
VanessaE |
AdBlock Plus works on firefox and chrom[e|ium] and does the job well |
| 14:22 |
rubenwardy |
+! |
| 14:22 |
rubenwardy |
*1 |
| 14:22 |
VanessaE |
I only have like, 2 sites in my whitelist (and only because they don't work without allowing their ads) |
| 14:23 |
VanessaE |
</offtopic> |
| 14:27 |
Taoki |
Anchakor: I can't imagine a good hosting service without ads |
| 14:28 |
Taoki |
But yeah, I use an ad blocker too... very happy I do :) |
| 14:29 |
Anchakor |
for some reason adblock plus doesn't work in chromium incognito mode |
| 14:34 |
Anchakor |
oh found an option to enable that |
| 14:35 |
rubenwardy |
Anchakor: drop box |
| 14:36 |
rubenwardy |
Any comments on Add item eat callback? ( https://github.com/minetest/minetest/pull/1255 ) |
| 14:37 |
|
CraigyDavi_ joined #minetest-dev |
| 14:39 |
Taoki |
rubenwardy: Don't see what that's useful for |
| 14:39 |
Taoki |
The way you can do actions in on_rightclick now works fine |
| 14:39 |
rubenwardy |
The modders obviously think so |
| 14:40 |
rubenwardy |
It allows hunger mods to override default behaviour without having a list of foods to override. |
| 14:40 |
rubenwardy |
eg, https://github.com/BlockMen/hud/blob/master/hunger.lua#L47 |
| 14:41 |
Taoki |
It feels like things that can be done well in Lua are being hardcoded without as much reason IMO. But dunno what to say otherwise |
| 14:41 |
rubenwardy |
It is in lua, and it actually gets rid of a hard coded thing |
| 14:42 |
rubenwardy |
Currently mods have to overwrite on_use (not on_rightclick) in node defs, which means they need a list of the foods to overwrite. |
| 14:43 |
Taoki |
I thought you just make on_rightclick() remove the food item while adding health |
| 14:43 |
|
sapier joined #minetest-dev |
| 14:43 |
rubenwardy |
on_rightclick is for placed nodes |
| 14:43 |
rubenwardy |
on_use is things in inventory |
| 14:43 |
Taoki |
Ah, sorry about that. Yeah on_use |
| 14:43 |
Taoki |
Still, there's just two actions. Adding health to the player, and decreasing the item count |
| 14:44 |
rubenwardy |
Can you please read what I say. |
| 14:44 |
sapier |
radio streaming in a game engine ... if it only wouldn't have been a serious suggestion ... :-/ |
| 14:45 |
Taoki |
I am, but apparently we're not understanding this well |
| 14:45 |
Taoki |
Or I'm missing an important part |
| 14:45 |
rubenwardy |
It would make no sense if you want to add a sound to your food in your mod. |
| 14:45 |
Taoki |
I already proposed a patch which does that |
| 14:45 |
Taoki |
Longer ago |
| 14:45 |
sapier |
rubenwardy what's special benefit for on_eat callback compared to on use? |
| 14:46 |
rubenwardy |
But it makes sense for hunger mods etc, as you are adding behaviour to another mods food |
| 14:46 |
* Taoki |
wonders the same thing as sapier |
| 14:46 |
rubenwardy |
sapier: you don't need to edit the item definition |
| 14:47 |
rubenwardy |
it is minetest.register_on_item_eat, not nodedef.item_eat |
| 14:47 |
sapier |
so you suggest adding a specialized variant of an existing callback to save a couple of lines only? |
| 14:47 |
Taoki |
Usually, if you define a food item you add its behavior in definition |
| 14:47 |
rubenwardy |
no existing callback |
| 14:48 |
sapier |
on_use does exist (or something doing this called quite similar) |
| 14:48 |
rubenwardy |
no. minetest.register_on_use does not exist. nodedef.on_use exists |
| 14:48 |
sapier |
and? |
| 14:48 |
rubenwardy |
This patch is like minetest.register_on_place_node for nodedef.on_place |
| 14:48 |
rubenwardy |
<rubenwardy> It allows hunger mods to override default behaviour without having a list of foods to override. |
| 14:49 |
sapier |
at price of causing quite some confusion |
| 14:50 |
sapier |
who is supposed to be called first node definition or the one you suggest? |
| 14:50 |
Taoki |
To me it feels like duplication for a small purpose |
| 14:50 |
rubenwardy |
both |
| 14:50 |
rubenwardy |
on_use = minetest.item_eat |
| 14:50 |
sapier |
there's no way to be both called first by definition |
| 14:50 |
rubenwardy |
minetest.item_eat then has a call back |
| 14:50 |
rubenwardy |
OMFG |
| 14:51 |
* Taoki |
scratches head... still doesn't really understand why this would be very hepuful, sorry :( |
| 14:51 |
rubenwardy |
You either use minetest.item_eat as your on_use, or you make your own on_use |
| 14:52 |
Taoki |
That sounds like duplicating the same thing |
| 14:52 |
sapier |
imho item_eat is a builtin example it's not to be used to call callbacks |
| 14:53 |
rubenwardy |
The alternative is to allow item_eat to be overridden |
| 14:53 |
sapier |
I don't remember this to be forbidden? |
| 14:54 |
rubenwardy |
minetest.item_eat(10) returns a function that is assigned to on_use. When that function is called, it runs the callbacks so the default eat behaviour can be overridden. |
| 14:55 |
rubenwardy |
What 'similar functionality' are you talking about? |
| 14:56 |
sapier |
by now I understand you don't wanna reimplement item_eat in your mod and therefore want core to provide another layer of indirection beeing quite specific to your mod. I can't imagine any other mod this feature might be usefull, tell me if you know about any |
| 14:56 |
rubenwardy |
sapier: oops |
| 14:56 |
rubenwardy |
<sapier> who is supposed to be called first node definition or the one you suggest? |
| 14:57 |
rubenwardy |
nodedef.on_use is called first, and the callbacks is in that |
| 14:57 |
sapier |
ok I see I have to look precisely what code does |
| 14:58 |
rubenwardy |
You would need to load before default to override minetest.item_eat, currently. |
| 15:01 |
sapier |
first of all we need to find out why it was done this way, this may be either an accident or a performance optimization. In later situation adding your callback loop would be nonsense |
| 15:01 |
rubenwardy |
Note the code is only 10 lines more. |
| 15:01 |
sapier |
what do you wanna do in those callbacks rubenwardy? |
| 15:02 |
|
Jordach joined #minetest-dev |
| 15:02 |
rubenwardy |
hunger is the main use |
| 15:02 |
rubenwardy |
immersive sounds could use it as well |
| 15:04 |
sapier |
ok I guess I now understood what you wanna do. But I'd not do it this way but do it like other callbacks are done, e.g. on_globalstep |
| 15:05 |
sapier |
that way it'd be consistent code and wouldn't add an additional callback mechanism |
| 15:06 |
sapier |
the question is how consistent are other callbacks right now |
| 15:06 |
Taoki |
sapier: https://github.com/minetest/minetest/pull/1266 Give this a merge please. I corrected my directional fog to work with the new tinted sun / moon (when gradients are present) |
| 15:08 |
rubenwardy |
Here is on alternate method, sapier, Taoki: https://github.com/rubenwardy/minetest/tree/item_eat_override |
| 15:08 |
rubenwardy |
wait... |
| 15:09 |
sapier |
where exactly rubenwardy ;-) |
| 15:09 |
rubenwardy |
done |
| 15:09 |
rubenwardy |
pushed |
| 15:09 |
rubenwardy |
https://github.com/rubenwardy/minetest/commit/965ec93cffbdce1d29fef92bf1e52c3a320592a4 |
| 15:10 |
* Taoki |
still doesn't fully get this, so best to conclude that I'm neither for nor (completely) against it |
| 15:10 |
sapier |
that's even more inconsistent ;) |
| 15:11 |
rubenwardy |
sapier: there is no on_globalstep |
| 15:11 |
sapier |
can you give an example within current code where callback loop is done in lua? |
| 15:12 |
sapier |
it is, right now callback tables are traversed by c++ code ... I'm not sure if this is best way to do it, but it's used troughout code |
| 15:12 |
rubenwardy |
https://github.com/minetest/minetest/blob/master/builtin/item.lua#L293 |
| 15:12 |
rubenwardy |
register_on_placenode |
| 15:12 |
sapier |
ok so obviously we already are inconsistent |
| 15:13 |
rubenwardy |
on node dig too |
| 15:13 |
sapier |
I suggest following the naming used there |
| 15:13 |
rubenwardy |
I have |
| 15:13 |
rubenwardy |
oh, in the for loop |
| 15:13 |
sapier |
yes |
| 15:14 |
sapier |
and of course check for return values being same as before |
| 15:14 |
sapier |
return value types and number of values of course |
| 15:14 |
sapier |
content can be different |
| 15:18 |
sapier |
taoki I'm gonna check it later I don't have time to merge something right now anyway |
| 15:18 |
Taoki |
ok |
| 15:19 |
sfan5 |
huh? who highlighted me here? |
| 15:26 |
|
Calinou joined #minetest-dev |
| 15:27 |
|
sapier left #minetest-dev |
| 15:43 |
|
Shardvexz joined #minetest-dev |
| 16:23 |
Megaf |
lots of folks are getting this now a days 13:09:27 markveidemanis | ERROR[ConnectionSend]: con(6/1)RunTimeouts(): Peer 2 has timed out. (source=peer->timeout_counter) |
| 16:27 |
Megaf |
PenguinDad: Yep, leaving the world in single player will cause that indeed |
| 16:28 |
Megaf |
13:27:20 PenguinDad | play singleplayer in latest git and look at the console when you leave the world with "exit to os" |
| 16:43 |
Calinou |
idea: split client and server configurations into two separate .conf files |
| 16:43 |
Calinou |
(just an idea I had, to clean up stuff) |
| 16:43 |
Calinou |
(an unique minetest.conf can get rather huge and confusing) |
| 16:43 |
Calinou |
would be useful for people who host server on their client PCs |
| 16:46 |
sfan5 |
Calinou: --config mtserver.conf , problem solved |
| 16:54 |
Calinou |
I mean by default |
| 16:54 |
Calinou |
obviously |
| 16:54 |
Calinou |
use different file unless specified otherwise |
| 17:00 |
|
OldCoder joined #minetest-dev |
| 17:19 |
|
kaeza joined #minetest-dev |
| 18:17 |
|
proller joined #minetest-dev |
| 18:18 |
Megaf |
+1 |
| 18:24 |
|
sapier joined #minetest-dev |
| 18:25 |
sapier |
Megaf that message is due to celeron not wanting to wait for a clean shutdown in singleplayer ;) |
| 18:26 |
sapier |
"singleplayer" is force timeouted in order to get fast exit |
| 18:28 |
|
specing left #minetest-dev |
| 18:28 |
sapier |
Calinou if you split it you add another class of error, either you need to extend api to fetch config from a specific file or merge them together, ... well merged variant doesn't work as you schould be able to set settings file specific too |
| 18:31 |
Calinou |
ah |
| 18:34 |
sapier |
I see settings example file getting a little bit big too but I guess almost noone really has all those settings set |
| 18:35 |
sapier |
so at least to me the benefit of splitting doesn't equalize the additional problems that may be caused |
| 18:40 |
|
us`0gb joined #minetest-dev |
| 18:58 |
|
CraigyDavi joined #minetest-dev |
| 19:02 |
|
us`0gb joined #minetest-dev |
| 19:12 |
|
daswort joined #minetest-dev |
| 19:16 |
Megaf |
ok |
| 19:16 |
Megaf |
so let´s keep it how it is at the moment |
| 19:17 |
sapier |
well the only way to remove it (completely) is disable logging ;-) |
| 19:18 |
RealBadAngel |
btw, disabling logging speeds up the game |
| 19:19 |
sapier |
of course ;-) |
| 19:19 |
Taoki |
heh... is logging enabled by default? |
| 19:19 |
sapier |
only errors and actions so most likely impact will be minor |
| 19:19 |
Taoki |
Maybe I should get rid of it then, until I need it |
| 19:19 |
Taoki |
Ah, then it should be fine |
| 19:20 |
Taoki |
If it visibly speeds up the engine, it likely means it logs several things each second or something |
| 19:20 |
RealBadAngel |
imho stable releases should go with logging turned off |
| 19:20 |
sapier |
why do you wanna hide errors? |
| 19:20 |
RealBadAngel |
if everythin runs fine, why slow the engine down? |
| 19:21 |
sapier |
I don't have numbers but I don't think it's gonna be a noticable difference |
| 19:21 |
RealBadAngel |
last time i checked that it was very noticeable |
| 19:21 |
sapier |
if everything runs fine you wont slow down anything because there wont be log entries |
| 19:21 |
proller |
sapier, you cant understand logging problem while you not run public server |
| 19:22 |
sapier |
well true I can't understand why someone setting up a server isn't capable of changeing loglevel to 0 |
| 19:22 |
RealBadAngel |
in case something is going wrong, allow it easily to be turned on again |
| 19:22 |
Taoki |
RealBadAngel: Did you see my directional fog correction yet? Tested it and it works fine with your tonemaps now :) |
| 19:22 |
RealBadAngel |
maybe an option in settings> |
| 19:23 |
sapier |
I don't believe in disabling something that doesn't happen |
| 19:23 |
RealBadAngel |
im telling you that this has real impact on game performance |
| 19:23 |
RealBadAngel |
and is sometimes a bottleneck |
| 19:24 |
sapier |
can you explain to me how something not happening can have an impact on performance? |
| 19:25 |
sapier |
action stream is user interaction only I can't believe users gonna cause actions as fast as having an impact |
| 19:25 |
RealBadAngel |
can you explain how the process that continously is writing "player xxx done something" and saving that to hdd could not have any impact on performance of the server? |
| 19:26 |
sapier |
usually modern os are quite good at caching, assuming we don't open logfile in sync mode, of course that's to verify |
| 19:26 |
RealBadAngel |
just turn it off and watch your fps :P |
| 19:27 |
CiaranG |
Totally anecdotal, but my server is logging > 2.5GB a day, and it has no noticable impact on performance if I don't do that |
| 19:27 |
sapier |
what loglevel do you have set CiaranG? |
| 19:27 |
CiaranG |
no idea |
| 19:27 |
CiaranG |
Most of that 2.5GB is coming from my mods, not the core |
| 19:27 |
sapier |
RealBadAngel: I'm gonna try, but if you're right I'd consider disabling log to be a workaround for a real bug |
| 19:28 |
CiaranG |
(actually, I have INFO) |
| 19:28 |
RealBadAngel |
disabling the logs by default were already a subject for a patch |
| 19:28 |
RealBadAngel |
quite a while ago |
| 19:28 |
RealBadAngel |
im not sure why the idea was lost |
| 19:29 |
sapier |
well CiaranG don't blame minetest if you change defaults to excessive logging ;-) |
| 19:29 |
CiaranG |
I'm not blaming anything, I'm logging that stuff on purpose |
| 19:29 |
CiaranG |
I'm just pointing out that it has no performance impact |
| 19:29 |
Calinou |
log_level 2 is nice for servers, keep logs of actions |
| 19:29 |
Calinou |
log_level 1 is more suited for clients, or even 0 |
| 19:30 |
|
diemartin joined #minetest-dev |
| 19:32 |
sapier |
RealBadAngel: at least on singleplayer I don't have any fps difference |
| 19:33 |
sapier |
debug_log_level = 2 is default, even less logging is 1 ... I don't recommend less then 1 because 1 should be errors only |
| 19:34 |
RealBadAngel |
sapier on my box it was like 5-6fps more |
| 19:34 |
Taoki |
sapier: How much does num_emerge_threads > 1 increase performance? |
| 19:34 |
sapier |
singleplayer with default levels? |
| 19:34 |
Taoki |
I have 8 threads for my CPU... can I set that to 8? |
| 19:35 |
sapier |
in best case it's gonna improve map generation speed |
| 19:35 |
Taoki |
ok. Sounds good too |
| 19:35 |
sapier |
but only if your cpu isn't already maxxed out |
| 19:35 |
sapier |
in this case it's gonna slow it down |
| 19:46 |
|
EvergreenTree joined #minetest-dev |
| 19:48 |
|
cg72 joined #minetest-dev |
| 19:54 |
|
CraigyDavi joined #minetest-dev |
| 19:56 |
Calinou |
it uses several threads for map generation and a few other stuff |
| 19:56 |
Calinou |
sapier, can caves still be buggy with multiple threads? |
| 19:56 |
Calinou |
or was this fixed? |
| 19:56 |
sapier |
they can be buggy with a single thread too but for more details you should ask hmmmm |
| 19:59 |
|
us`0gb joined #minetest-dev |
| 20:16 |
* Megaf |
will setup to identical servers to test that log thing |
| 20:18 |
|
us`0gb joined #minetest-dev |
| 20:18 |
proller |
stop count performance with fps |
| 20:19 |
sapier |
especially for servers |
| 20:19 |
* Megaf |
limited his client FPS to 24 |
| 20:19 |
Megaf |
really, theres no need for more than that |
| 20:20 |
Megaf |
minumum and max fps set to 24 fps, so the menu fps |
| 20:20 |
sapier |
megaf you don't need to test log level 4 against 0 that's gonna have an effect for sure, try 1 or 2 against 0 |
| 20:20 |
Megaf |
also fixed minimum and maximum view rang to 128 |
| 20:20 |
Megaf |
sapier: I will setup and run two server, then you folks can join and thest yourselves |
| 20:20 |
sapier |
1 schouldn't have any effect while 2 most likely has a minor effect but not within range to measure with fps |
| 20:20 |
Megaf |
test* |
| 20:21 |
Megaf |
but it will take a while |
| 20:21 |
Megaf |
6 hours at least |
| 20:23 |
VanessaE |
Megaf: your machine will fall over if you go to my Survival or Creative server with those settings |
| 20:24 |
Megaf |
so you should optimize your server |
| 20:26 |
VanessaE |
it's not the server. |
| 20:26 |
VanessaE |
it's a client issue. |
| 20:26 |
VanessaE |
problems in the rendering pipeline |
| 20:32 |
|
werwerwer joined #minetest-dev |
| 20:40 |
Megaf |
my server makes some clients suffer because of lights, lots of lights poles and torches |
| 20:40 |
proller |
i like 3000 view range ;) |
| 20:40 |
sapier |
light calculation is one of most time consuming tasks in minetest |
| 20:41 |
sapier |
it can probably improved but you can't realy fix it for current lighting mechanism |
| 20:44 |
Megaf |
sapier: and theres not much to calculate... no reflections, no shadows.. |
| 20:44 |
Megaf |
no refraction |
| 20:45 |
sapier |
but light propagation is done manually |
| 20:45 |
Megaf |
the GPU should do that |
| 20:45 |
sapier |
not in minetest |
| 20:46 |
sapier |
that's why I said it can't be fixed for current lighting mechanism |
| 20:46 |
|
tomreyn joined #minetest-dev |
| 20:46 |
sapier |
I agree to you but that's a mayor change in engine, by now noone tried to do that |
| 20:47 |
* Megaf |
looks at RealBadAngel |
| 20:51 |
|
PenguinDad joined #minetest-dev |
| 20:55 |
|
us`0gb joined #minetest-dev |
| 20:58 |
|
smoke_fumus joined #minetest-dev |
| 21:06 |
RealBadAngel |
server is going to calculate only light level at the node |
| 21:06 |
RealBadAngel |
all other stuff will be removed from server's code |
| 21:07 |
RealBadAngel |
and will be passed onto irrlicht engine |
| 21:07 |
proller |
shut up and commit!!!! ;) |
| 21:07 |
RealBadAngel |
thats not a change that will happen with one commit just |
| 21:08 |
RealBadAngel |
but shaders are already pretty ready for light sources |
| 21:08 |
sapier |
RealBadAngel: I'd start with enabling client to do light calculation on it's own while keeping server lighting for compatibility reasons |
| 21:09 |
RealBadAngel |
https://github.com/RealBadAngel/minetest/blob/shaders/client/shaders/nodes_shader/opengl_vertex.glsl#L84 |
| 21:09 |
RealBadAngel |
shaders do already use one fixed light source |
| 21:09 |
proller |
sapier, not you!!! |
| 21:09 |
sapier |
you have to explayn what this is supposed to do |
| 21:09 |
RealBadAngel |
this? |
| 21:09 |
sapier |
why so frightened proller ? |
| 21:09 |
RealBadAngel |
it is SUN |
| 21:10 |
proller |
we loosing ugly lighting 8( |
| 21:10 |
sapier |
btw did you find out why your client is broken that way? |
| 21:10 |
sapier |
what's your download performance? can you tell the real numbers |
| 21:10 |
sapier |
? |
| 21:11 |
proller |
me? |
| 21:11 |
sapier |
well you said downloading is only a few kb on singleplayer, for me it's multiple mb so either you just did bashing again or something is different on your setup |
| 21:12 |
sapier |
it's multiple mb for everyone else who provided feedback too |
| 21:13 |
proller |
wait |
| 21:13 |
sapier |
it'd not be first time for a bug to occur for a single person only but without that persons help it's next to impossible to fix it |
| 21:18 |
proller |
ht858q |
| 21:19 |
proller |
http://paste.org.ru/?ht858q |
| 21:19 |
proller |
NeuromancerGame login |
| 21:19 |
proller |
netstat -w1 -I lo0 |
| 21:19 |
proller |
speed unstable |
| 21:20 |
sapier |
that doesn't tell anything without additional information |
| 21:20 |
proller |
and fuckin low for local game |
| 21:20 |
proller |
old tuned connection give me stable 3-4-5 mbps |
| 21:20 |
sapier |
stable transmission or stable load ? that's a big difference |
| 21:21 |
proller |
transmission |
| 21:21 |
sapier |
proven by what netstat? |
| 21:21 |
proller |
?? |
| 21:21 |
sapier |
netstat does only show load |
| 21:21 |
proller |
what? |
| 21:21 |
sapier |
netstat doesn't show real data successfully transmitted but only what was pumped out |
| 21:22 |
proller |
what??? |
| 21:22 |
proller |
i can show you freeminer/next output |
| 21:22 |
sapier |
I'll try again ... if some udp server sends 1000mb but client will only be able to handle 100mb, load will be 1000mb but transmission only 100mb |
| 21:23 |
proller |
and lot of errs |
| 21:23 |
sapier |
if same server transmitts 200mb but client is capable of receiving 150 mb ...load is 200mb while transmission is 150 mb |
| 21:23 |
proller |
STOP THIS SHIT NOW |
| 21:23 |
sapier |
in this example later one is what you want ... but netstat only tells about load |
| 21:25 |
sapier |
proller I wanna help you, I believe you feel it to be slower but none of the data provides really proves this, it doesn't prove the opposit either |
| 21:27 |
sapier |
I cannot reproduce your issue any test I did shows up new implementation beeing rock solid and way faster for texture transmission. If you can provide a testcase where this isn't true I can check why this happens. |
| 21:35 |
proller |
http://paste.org.ru/?c6h9g9 - enet |
| 21:35 |
proller |
2 seconds transfer |
| 21:35 |
sapier |
I don't compare to enet |
| 21:36 |
sapier |
as well as I don't compare enet to tcp which most likely would be 0.2s for your case |
| 21:37 |
proller |
you can compare your shit only with full shit |
| 21:37 |
proller |
also http transfer will be much slower |
| 21:38 |
proller |
(in current implementation) |
| 21:38 |
sapier |
provide a testcase and I'm gonna fix it |
| 21:38 |
proller |
fix what? |
| 21:38 |
sapier |
not some random numbers that could or couldn't tell something |
| 21:38 |
proller |
bye |
| 21:40 |
Megaf |
sapier: about the log thing again |
| 21:40 |
Megaf |
it's not the amount of CPU power it uses/requires |
| 21:40 |
Megaf |
and so so much about disk IO |
| 21:40 |
sapier |
yes |
| 21:40 |
Megaf |
the CPU can only do one thing at time |
| 21:41 |
Megaf |
so, or it prints logs or processes player action or write something to somewhere |
| 21:41 |
Megaf |
and here is the problem |
| 21:41 |
Megaf |
perhaps we could delay the log or make it go in batches |
| 21:41 |
sapier |
but as long as log activity stays withing a certain threshold I'd expect a modern os to handle this (for non sync file operation) without noticable impact to calling application |
| 21:42 |
Megaf |
sapier: but it causes a lot of impact |
| 21:42 |
sapier |
logfile size may have an impact too (at least to me) |
| 21:45 |
|
grrk-bzzt joined #minetest-dev |
| 22:07 |
VanessaE |
Megaf: if the CPU isn't writing that much to disk at once, the OS will already gather it in batches anyway |
| 22:08 |
|
lemon joined #minetest-dev |
| 22:08 |
VanessaE |
it'll cache the write into memory, outside of Minetest's control, and shuffle it off to disk when it gets a chance, unless Minetest expressly tells the OS "no, write it to disk NOW" |
| 22:08 |
|
Weedy_ joined #minetest-dev |
| 22:09 |
VanessaE |
what I would be more worried about, from a logging standpoint, isn't the bandwidth, so much as the storage requirement. if your storage is limited (e.g. a VPS), you have to keep a close eye on the size of your logs. it's not hard for a log file to grow into the few-gigs range. |
| 22:10 |
Megaf |
VanessaE: we can tell the same about rollback.sqlite |
| 22:10 |
|
SmugLeaf joined #minetest-dev |
| 22:11 |
VanessaE |
yeah but rollback is one you definitely want to keep |
| 22:11 |
VanessaE |
bbl |
| 22:13 |
sapier |
having rollback and logfile on same disk may (in worst case) cause even more lag then summ of log and rollback |
| 22:14 |
sapier |
e.g. if those are done the very same moment and need to be written to completely different locations on disk |
| 22:14 |
|
EvergreenTree joined #minetest-dev |
| 22:14 |
sapier |
at least theoretically ... in practice os caching may even equalize this effect |
| 22:34 |
|
harrison joined #minetest-dev |
| 23:04 |
|
kaeza joined #minetest-dev |
| 23:36 |
|
kaeza joined #minetest-dev |
