| Time |
Nick |
Message |
| 00:32 |
|
Noideawhatmedoin joined #minetest |
| 00:33 |
Noideawhatmedoin |
I'm officially lost, after literally a week of research I still can't "render distance" in Minetest, please help. |
| 00:35 |
specing |
- + |
| 00:44 |
Noideawhatmedoin |
(I know about that, I mean how do I actually make the machine render more, even if I set it to 4000 and enable unlimited viewing distance. It still only renders around a radius of 400 or so) |
| 00:50 |
specing |
Well |
| 00:50 |
specing |
you probably don't have those blocks loaded |
| 00:50 |
specing |
So, the question becomes "how do you force load 4000 blocks ahead and bring your machine to a crawl?" |
| 00:57 |
Noideawhatmedoin |
Teach me the ways of the kernel panic inducing |
| 01:08 |
Noideawhatmedoin |
(This is a total guess, but "overflows" seem to be a recurring theme in Minetest Bugs. Any chance it's that?) |
| 01:09 |
Noideawhatmedoin |
(The render distance (despite relentless tweaking) seems to happen on both my cellphone and my main system.) |
| 01:15 |
|
Taoki joined #minetest |
| 01:27 |
Noideawhatmedoin |
Welp... |
| 01:27 |
Noideawhatmedoin |
How do you force load 4000 blocks ahead and bring your machine to a crawl? |
| 01:34 |
Noideawhatmedoin |
(Not to mention I actually plan on getting a legit Desktop soon, so it would be nice so for when I decide to bring it "to a crawl"... |
| 01:34 |
Noideawhatmedoin |
) |
| 02:03 |
|
sagax joined #minetest |
| 03:22 |
Noideawhatmedoin |
Say I actually wanted to force Minetest to churn out more blocks to the screen at once, regardless of the performance implications, how would I go about doing this? |
| 03:26 |
MTDiscord |
<Jordach> what are the + and - buttons on your keyboard for? |
| 03:26 |
MTDiscord |
<Jordach> you can change viewing distance in game |
| 03:26 |
MTDiscord |
<Jordach> also, the camera "clips" or stops rendering things after a certain distance |
| 03:55 |
MTDiscord |
<exe_virus> You can change the default load distance as well as the render distance, so your computer/server will start generating and sending you chunks well over 2000 nodes away |
| 04:02 |
|
Taoki joined #minetest |
| 04:25 |
|
fattywompus_ joined #minetest |
| 04:37 |
|
dabbill joined #minetest |
| 04:43 |
|
awell joined #minetest |
| 04:56 |
|
Verticen joined #minetest |
| 05:00 |
|
MTDiscord joined #minetest |
| 06:45 |
|
MDude joined #minetest |
| 06:57 |
|
Follpvosten[m] joined #minetest |
| 06:57 |
Follpvosten[m] |
Pretty sure you can force blocks to load with /emergeblocks |
| 07:29 |
|
Peppy joined #minetest |
| 07:31 |
|
Noideawhatmedoin joined #minetest |
| 07:36 |
|
aheinecke joined #minetest |
| 07:42 |
|
Flabb joined #minetest |
| 07:56 |
Noideawhatmedoin |
Doesn't emergeblocks just generate new blocks? |
| 08:00 |
|
ShadowNinja joined #minetest |
| 08:03 |
Noideawhatmedoin |
(Also using /emergeblocks doesn't really seem to do anything) |
| 08:03 |
Noideawhatmedoin |
(At least with respect to render distance) |
| 08:05 |
Emerald2 |
Doesn't it just run mapgen again for the given area? |
| 08:08 |
Noideawhatmedoin |
(My understanding is it tries to generate new blocks within the specified area (whether it's in a certain range or two specific points.) |
| 08:19 |
Noideawhatmedoin |
After doing an emergeblocks, flying around does load in blocks (as expected), however nothing about the rate at which blocks appear hasn't really changed. |
| 08:23 |
|
Noideawhatmedoin joined #minetest |
| 08:40 |
|
Fleckenstein joined #minetest |
| 08:40 |
|
Fleckenstein left #minetest |
| 08:58 |
MTDiscord |
<IhrFussel> No guys there IS a hardcoded render distance and no matter what you do, you cannot increase it beyond that without modifying the c++ code |
| 09:00 |
MTDiscord |
<IhrFussel> They were probably asking what they need to change to make it render even further...such questions would be better suited for dev (maybe they already asked there) |
| 09:01 |
|
TomTom joined #minetest |
| 09:11 |
|
FeXoR joined #minetest |
| 09:13 |
|
calcul0n joined #minetest |
| 10:23 |
|
Ruud joined #minetest |
| 10:35 |
|
proller joined #minetest |
| 10:57 |
celeron55 |
max_block_send_distance and max_block_generate_distance certainly have to be set appropriately, and maybe something else |
| 10:58 |
celeron55 |
you should probably pastebin your minetest.conf |
| 11:26 |
|
Wuzzy joined #minetest |
| 11:35 |
rubenwardy |
There isn't a hardcoded render distance IhrFussel, you just need to increase the above settings |
| 12:02 |
|
Fixer joined #minetest |
| 12:52 |
|
Peppy joined #minetest |
| 13:04 |
|
Fixer_ joined #minetest |
| 13:10 |
MTDiscord |
<IhrFussel> I'm pretty sure someone in the forum once complained that no matter what they do they couldn't get it to render past 400 or so nodes...or maybe it was much more...but they basically tried all possible settings that are related |
| 13:26 |
celeron55 |
well there is some limit due to the z buffer settings you have to give to the camera |
| 13:27 |
celeron55 |
the shorter you make the z buffer range, the more accurate the z buffer will be, so there's reason to not set it stupidly far |
| 13:27 |
celeron55 |
altough it could be dynamically adjusted |
| 13:27 |
celeron55 |
(maybe it is? i don't remember) |
| 13:30 |
celeron55 |
https://github.com/minetest/minetest/blob/master/src/client/camera.cpp#L625 |
| 13:30 |
Noideawhatmedoin |
Greetings again, I didn't think I'd get the attention of the name at the top of the list. Anyway, there's actually two "minetest.confs" one on my server and then of course the one on the client system. |
| 13:30 |
celeron55 |
well that's some additional limit |
| 13:30 |
celeron55 |
looks like the viewing range setting is clamped to 400 nodes |
| 13:31 |
Noideawhatmedoin |
(Should I post both?) |
| 13:31 |
celeron55 |
maybe |
| 13:31 |
celeron55 |
not sure if that's in nodes or in view units |
| 13:32 |
celeron55 |
m_cameranode->setFarValue() is given a dynamic value up to 2000 nodes so it's not the problem like i suspected |
| 13:32 |
celeron55 |
IhrFussel: try to adjust that line and see if it helps https://github.com/minetest/minetest/blob/master/src/client/camera.cpp#L637 |
| 13:33 |
celeron55 |
hmm wait no, that is in nodes |
| 13:34 |
celeron55 |
so it can't be the limiting factor |
| 13:35 |
celeron55 |
https://forum.minetest.net/viewtopic.php?p=315589&sid=5ebc84cf8a07a0e2df41128c1506f5e3#p315589 |
| 13:35 |
celeron55 |
apparently block_send_optimize_distance and client_mapblock_limit have to be also adjusted |
| 13:36 |
celeron55 |
the system is veeery inefficient in transferring relevant blocks to the client at large ranges though |
| 13:36 |
celeron55 |
as there's no height limit |
| 13:37 |
celeron55 |
and it won't guess what to send |
| 13:37 |
celeron55 |
it used to do that, and that created weird voids in oceans and people didn't like it either |
| 13:37 |
celeron55 |
so it was removed |
| 13:38 |
Noideawhatmedoin |
(Could it be an overflow bug, that seems to be a weird trend I noticed among some bug reports (Aside from the hardcoded limits that incite some form of request to remove/raise the limits *cough cough*World height*cough cough*?) |
| 13:46 |
Noideawhatmedoin |
(I'm not sure if this counts as the same issue, but in the past I've tried to up some of the various "block send" variables and couldn't seem to really get past 40 or so. The world loads and is just sort of stuck in some kind of blank preload state without any evidence of the system slowing down or doing something else that is a result of rendering, so either there's some kind of weird limit.) |
| 14:34 |
|
I_am_6r1d joined #minetest |
| 14:35 |
|
kamdard joined #minetest |
| 14:45 |
|
nuala joined #minetest |
| 14:54 |
|
kamdard joined #minetest |
| 14:59 |
|
m42uko joined #minetest |
| 15:11 |
|
sec^nd joined #minetest |
| 15:14 |
tango_ |
I wonder if a possible improvements for the z buffer accuracy could be to do a split rendering, "far" vs "close" blocks, and then overlay the renders |
| 15:15 |
tango_ |
so you'd have one for view distance, say, 200 to 800 (low accuracy OK because blocks are far away anyway) vs another for 1 to 200 |
| 15:15 |
tango_ |
this of course doesn't solve the inefficiency in transferring blocks |
| 15:16 |
tango_ |
celeron55: couldn't a height limit be computed based on the view angle? |
| 15:16 |
tango_ |
or is there again an accuracy issue for far away blocks? |
| 15:26 |
|
Verticen joined #minetest |
| 15:42 |
|
erlehmann joined #minetest |
| 15:45 |
Noideawhatmedoin |
Did this just turn into #minetest-dev? |
| 15:48 |
Noideawhatmedoin |
(Sorry was doing things, could I post a minetest.conf on a privacy friendly clone of Pastebin? |
| 15:48 |
Noideawhatmedoin |
) |
| 15:59 |
|
Verticen joined #minetest |
| 16:11 |
|
Lukwe joined #minetest |
| 16:18 |
|
Copenhagen_Bram joined #minetest |
| 16:18 |
|
lisac joined #minetest |
| 16:26 |
Noideawhatmedoin |
(Well actually I'm stupid I somehow figured things out) |
| 16:27 |
Noideawhatmedoin |
(Actually no wait scratch that I failed at this) |
| 16:28 |
tango_ |
Noideawhatmedoin: sprunge.us? |
| 16:28 |
Noideawhatmedoin |
Well actually I was going to use bin.snopyta.org |
| 16:30 |
Noideawhatmedoin |
(Never heard of this, looks cool I might use it at some point in the future) |
| 16:42 |
|
Peppy joined #minetest |
| 17:13 |
|
Flabb joined #minetest |
| 17:36 |
|
Talkless joined #minetest |
| 17:45 |
|
numzero joined #minetest |
| 18:04 |
|
turtleman joined #minetest |
| 18:27 |
|
absurb joined #minetest |
| 18:32 |
|
calcul0n_ joined #minetest |
| 18:35 |
|
fluxflux joined #minetest |
| 18:37 |
|
proller joined #minetest |
| 18:44 |
|
FeXoR joined #minetest |
| 18:47 |
|
homthack joined #minetest |
| 18:55 |
|
FeXoR joined #minetest |
| 19:04 |
|
schrodinger72 joined #minetest |
| 19:05 |
MinetestBot |
[git] SmallJoker -> minetest/minetest: Input: Fix on_rightclick called when placing into air 3176dae https://git.io/JIvYa (2020-11-29T18:20:45Z) |
| 19:41 |
|
daiNoZord joined #minetest |
| 20:13 |
|
Verticen joined #minetest |
| 20:17 |
|
Peppy joined #minetest |
| 20:23 |
|
mizux joined #minetest |
| 20:25 |
|
fleeky_ joined #minetest |
| 20:30 |
|
Hawk777 joined #minetest |
| 20:44 |
specing |
Why do people use seperate nodes for each state of growth? |
| 20:50 |
|
Andrey01 joined #minetest |
| 20:51 |
rubenwardy |
because you want a different appearance |
| 20:51 |
Andrey01 |
hello, is there a support for mesh wielded items currently? |
| 20:55 |
Andrey01 |
specing: because nodes are fully static, it is impossible to change their textures/models and etc |
| 20:55 |
rubenwardy |
yes, use a mesh node |
| 20:56 |
rubenwardy |
that was have a mesh wield item |
| 20:56 |
rubenwardy |
*Will |
| 20:57 |
Andrey01 |
nodes means they could be put, but I want them to be just items that couldn`t be out |
| 20:57 |
Andrey01 |
put* |
| 20:59 |
Andrey01 |
I don`t want to register a node |
| 21:00 |
specing |
Andrey01: rubenwardy: it's not possible to vary apperance via param2 or what was it? |
| 21:00 |
|
proller joined #minetest |
| 21:00 |
Andrey01 |
of course, no. 'param2' is used for representing an orientation of a node |
| 21:01 |
specing |
8 bits for orientation? |
| 21:01 |
Andrey01 |
this is quite other |
| 21:04 |
Andrey01 |
yeah, as I understand, there is no a support for just items, also it is pity this is closed: https://github.com/minetest/minetest/pull/8725 |
| 21:14 |
rubenwardy |
param2 does more than that |
| 21:15 |
tango_ |
param2 is also used for levels |
| 21:15 |
rubenwardy |
you can vary colors, for example |
| 21:15 |
rubenwardy |
and that |
| 21:15 |
tango_ |
so why couldn't it be used to map levels to texture? |
| 21:16 |
rubenwardy |
probably could be, but it hasn't |
| 21:17 |
rubenwardy |
Andrey01: you can override on_place to stop nodes from placing |
| 21:17 |
rubenwardy |
also, that PR was closed because the author left Minetest, not because it wasn't desirable |
| 21:17 |
rubenwardy |
(I'm aware they've left) |
| 21:49 |
|
erlehmann joined #minetest |
| 21:51 |
|
Noideawhatmedoin joined #minetest |
| 22:01 |
|
SwissalpS joined #minetest |
| 22:02 |
|
erlehmann joined #minetest |
| 23:06 |
|
Elouin9 joined #minetest |
| 23:07 |
|
testman joined #minetest |
| 23:12 |
|
sec^nd joined #minetest |
| 23:12 |
|
bigfoot547 joined #minetest |
| 23:15 |
|
est31 joined #minetest |
| 23:16 |
|
milkt joined #minetest |
| 23:54 |
|
nuala joined #minetest |
