| Time |
Nick |
Message |
| 00:09 |
|
sloantothebone joined #minetest-dev |
| 00:11 |
sloantothebone |
So, anyone planning to make a movement API to allow players to be pushed around by flowing water, walk slower on some kind of mud block, and be knocked back by swords and bows? |
| 00:12 |
|
pozzoni joined #minetest-dev |
| 00:31 |
Obani |
sloantothebone, do it |
| 00:42 |
sloantothebone |
Ok, can I see the code for movement? |
| 00:44 |
kaadmy |
good luck sloantothebone, afaik the movement is a huge mess |
| 01:28 |
VanessaE |
something I think Minetest needs to start doing is multi-threading the "initializing nodes" step, if possible. |
| 01:28 |
VanessaE |
for big games like dreambuilder, that step takes a good 45 seconds on slower boxes |
| 02:05 |
|
Etzos joined #minetest-dev |
| 02:18 |
|
Taoki joined #minetest-dev |
| 02:55 |
|
gregorycu joined #minetest-dev |
| 02:55 |
gregorycu |
Guess who's back |
| 02:56 |
VanessaE |
hey greg |
| 02:56 |
gregorycu |
How goes the release? |
| 02:58 |
|
ssieb joined #minetest-dev |
| 02:59 |
sofar |
gregorycu: could use your help ;) |
| 02:59 |
sofar |
maybe review some tagged milestone PRs? |
| 03:01 |
gregorycu |
I can review, not a core dev so my +1 counts for nothing :P |
| 03:02 |
sofar |
it'll help |
| 03:02 |
sofar |
or fix some remaining open bugs |
| 03:02 |
gregorycu |
Yeah, looking now |
| 03:07 |
sloantothebone |
So... who messed up the movement code?? |
| 03:09 |
VanessaE |
messed it up how? |
| 03:14 |
gregorycu |
I assume he means collision detection? That's nasty code if it's fucked up |
| 03:15 |
|
proller joined #minetest-dev |
| 03:27 |
gregorycu |
sloantothebone ? |
| 03:32 |
sloantothebone |
(02:42:44 AM) sloantothebone: Ok, can I see the code for movement? |
| 03:32 |
sloantothebone |
(02:44:33 AM) kaadmy: good luck sloantothebone, afaik the movement is a huge mess |
| 03:36 |
gregorycu |
Oh, you mean the code is bad |
| 03:36 |
gregorycu |
Yeah, it's a mess |
| 03:36 |
|
proller joined #minetest-dev |
| 03:37 |
gregorycu |
Probably just grew organically and nobody bothered to refactor it |
| 03:37 |
gregorycu |
(Refactoring such things are very hard - hard to test every use case) |
| 03:58 |
|
gregorycu joined #minetest-dev |
| 04:08 |
gregorycu |
Anyone know slemonide? |
| 04:08 |
gregorycu |
(GitHub account name) |
| 04:15 |
sofar |
yes |
| 04:15 |
sofar |
he's not on much |
| 04:15 |
sofar |
think his irc name is different |
| 04:19 |
gregorycu |
Also sofar |
| 04:19 |
gregorycu |
Did you get "elevated" to the status of core dev? |
| 04:19 |
sofar |
after this release |
| 04:20 |
gregorycu |
Well, after this release, congratulations! |
| 04:35 |
gregorycu |
This code makes baby Jesus cry |
| 05:18 |
VanessaE |
heh |
| 05:19 |
VanessaE |
as long as it doesn't open up black holes or result in the deaths of innocent kittens, that's okay. |
| 05:21 |
gregorycu |
Well, I'm not sure |
| 05:45 |
|
Hunterz joined #minetest-dev |
| 06:21 |
ssieb |
I found a python library to connect to a minetest server, but at this point it uses an older protocol version. Does current minetest reject anything below 25? |
| 06:22 |
sofar |
don't think so |
| 06:22 |
sofar |
should be easy to fixup anyway |
| 06:22 |
VanessaE |
servers can be configured to reject though |
| 06:22 |
VanessaE |
strict_protocol_checking |
| 06:24 |
ssieb |
I made sure that was set to false |
| 06:25 |
ssieb |
but the client is getting kicked with TOCLIENT_ACCESS_DENIED_LEGACY |
| 06:25 |
ssieb |
I'll probably not bother any more with getting it to work and just convert it to a current protocol version... |
| 06:38 |
|
proller joined #minetest-dev |
| 07:12 |
gregorycu |
What's the python lib for? |
| 07:15 |
ssieb |
gregorycu: It implements the minetest network protocol. |
| 07:15 |
|
fling joined #minetest-dev |
| 07:16 |
gregorycu |
I guessed as much, but to what end? |
| 07:16 |
gregorycu |
What do people use it for? |
| 07:17 |
ssieb |
I don't know if anyone uses it. :-) |
| 07:18 |
ssieb |
It could be useful for making a bot (the actual intended purpose of the project this library came from). Or just for testing the server. I 'm looking at using it for both purposes. |
| 07:20 |
gregorycu |
Sounds good, I like python |
| 07:21 |
gregorycu |
Not a fan of Lua? |
| 07:27 |
gregorycu |
I could actually use a bot right now, lol |
| 07:27 |
gregorycu |
Testing a memory leak |
| 07:28 |
ssieb |
lua's ok for mods, but I much prefer python for real coding |
| 07:28 |
gregorycu |
Good lad |
| 07:29 |
gregorycu |
I was contemplating at a point in time adding python support for mods |
| 07:29 |
gregorycu |
So you could write mods in python |
| 07:29 |
VanessaE |
not sure that would work out so well |
| 07:30 |
VanessaE |
python has a habit of breaking things between major revisions. |
| 07:30 |
gregorycu |
From which perspective? |
| 07:30 |
gregorycu |
Yes, major revisions |
| 07:30 |
gregorycu |
We wouldn't be tethered to the python release schedule in any case |
| 07:31 |
gregorycu |
We would dictate what version of python was to be used |
| 07:32 |
gregorycu |
This isn't a serious suggestion, was just going to be something I do off the radar |
| 07:32 |
gregorycu |
Are torches meant to work underground? |
| 07:32 |
VanessaE |
well if you locked it to a specific version then I'd say that's not so bad. |
| 07:32 |
VanessaE |
and yes, they are. |
| 07:33 |
gregorycu |
And at night? |
| 07:33 |
VanessaE |
yep of course |
| 07:33 |
gregorycu |
That's what I thought |
| 07:34 |
gregorycu |
Gotta love bugs |
| 07:36 |
gregorycu |
Man, open source coding sucks |
| 07:36 |
gregorycu |
I have questions, and i want answers now |
| 07:36 |
gregorycu |
Not wait until people drag themselves onto IRC |
| 07:38 |
VanessaE |
heh |
| 07:38 |
VanessaE |
could be worse |
| 07:38 |
VanessaE |
if it were closed-source you'd be forced to do even uglier things like send emails or make phone calls :) |
| 07:39 |
gregorycu |
I suppose my gripe is with distributed programming |
| 07:40 |
gregorycu |
(not a serious gripe, just joking of course) |
| 07:40 |
gregorycu |
thought it would be just dandy if Paramat were around |
| 07:41 |
VanessaE |
well that's the thing, he might very well be. |
| 07:41 |
VanessaE |
he reads the logs, probably has a scaper set to ping him if he's mentioned. |
| 07:43 |
gregorycu |
What is a mesecons lamp? |
| 07:43 |
VanessaE |
a small square light that turns on with a mesecons signal. |
| 07:43 |
VanessaE |
about a quarter the size of a homedecor thin glowlight |
| 07:44 |
VanessaE |
(the "duh" answers are always the best huh? :D ) |
| 07:44 |
gregorycu |
Sweet, torch has a comparable lighting effect I assume? |
| 07:44 |
|
nrzkt joined #minetest-dev |
| 07:45 |
VanessaE |
something similar yeah |
| 07:45 |
VanessaE |
I think HD's thin glowlight has the same light output as a torch |
| 07:45 |
VanessaE |
(or maybe one step higher) |
| 07:45 |
VanessaE |
but all light sources have the same effect. |
| 07:46 |
VanessaE |
they fall off within 16 nodes I guess, always the same pattern. it's just a matter of how bright the source is. |
| 07:47 |
gregorycu |
Excellent |
| 07:48 |
VanessaE |
(a pattern which, when I last checked, was not-quite-diamond-shaped, owing to the geometry of the mesh being lit. circular-shaped would be better of course) |
| 07:49 |
gregorycu |
Yeah, that's due to the manhatten distance being used for lighting calculations |
| 07:50 |
VanessaE |
RBA had planned to do real lighting without limitations like that. |
| 07:51 |
gregorycu |
It gets difficult if we only store integer light values against each node |
| 07:51 |
VanessaE |
how else would you do it, aside from scaled values maybe? |
| 07:52 |
gregorycu |
The reason you have a diamond pattern is that light spreads to its neighbours, but not diagonally |
| 07:52 |
VanessaE |
right |
| 07:53 |
gregorycu |
If it were to spread diagonally it would be better |
| 07:53 |
VanessaE |
it would certainly be brighter |
| 07:53 |
gregorycu |
Yes |
| 07:54 |
gregorycu |
But it would form a box like pattern |
| 07:54 |
gregorycu |
As opposed to a diamond |
| 07:54 |
VanessaE |
right - how hard would it be, really, to make it circular? |
| 07:54 |
gregorycu |
If it were to spread diagonally, you should make it reduce by sqrt(2) |
| 07:55 |
gregorycu |
The thing is, if you use integers everywhere, it would round |
| 07:55 |
gregorycu |
If you kept things as float, it would look proper |
| 07:56 |
gregorycu |
(if you have excel, you can probably try this out) |
| 07:57 |
VanessaE |
what's the performance and/or storage impact? (seems to me there wouldn't be one, as even the shortest int is generally 16 or 32 bit anyway, and one could use half-precision floats for thios maybe) |
| 07:57 |
VanessaE |
-o |
| 07:58 |
gregorycu |
double would be faster than float |
| 07:58 |
VanessaE |
(I don't have excel, but I think I have a good idea of the result) |
| 07:58 |
nore |
VanessaE: storage would be affected a lot |
| 07:59 |
VanessaE |
gregorycu: because of the 32-bit alignment? |
| 07:59 |
nore |
we currently only use 8 bits for light, 4 of which for day and 4 for night |
| 07:59 |
VanessaE |
nore: I figured. |
| 07:59 |
gregorycu |
nore, you can answer this question for me |
| 08:00 |
VanessaE |
wait, I thought it was 4 bits for sunlight, and 4 for artificial light (regardless of time-of-day)? |
| 08:00 |
gregorycu |
no |
| 08:00 |
nore |
VanessaE: I wished it was so |
| 08:00 |
gregorycu |
if a placed schematic has correct lighting during the day, but not during night, what didn't happen? |
| 08:00 |
nore |
and when light changes, I hope this will be changed too |
| 08:00 |
VanessaE |
bleh, that's really a terrible design :-/ |
| 08:01 |
nore |
that means the night lighting bits are incorrect |
| 08:01 |
VanessaE |
there's really no reason for there to be anything more than one single light value there, since you already know the time of day (and hence sun angle, were it useful) |
| 08:01 |
nore |
VanessaE: yes, but you don't know if sunlight is visible or not |
| 08:01 |
nore |
so it makes sense |
| 08:01 |
VanessaE |
mmmh |
| 08:02 |
nore |
however, it would be a lot better to put a value for sunlight and one for artificial light |
| 08:02 |
VanessaE |
so use one bit for visible-or-not and the other 7 for the light. |
| 08:02 |
nore |
since this would give more information on the node (i.e. whether it is exposed to sunlight) |
| 08:02 |
gregorycu |
Also, the question remains if we need to be so frugal with out bits |
| 08:03 |
gregorycu |
But this is a more general question |
| 08:03 |
nore |
VanessaE: the problem with that would be a sharp cut at the bottom of a mine hole for example |
| 08:03 |
VanessaE |
honestly, no we don't. the largest map I've ever heard of was Xanadu, prior to the recent reset, and that was only 30 GB in sqlite format. |
| 08:03 |
nore |
(nodes at 15 light next to ones at 0) |
| 08:04 |
gregorycu |
Isn't that somewhat realistic? |
| 08:04 |
nore |
keep in mind that a node currently only takes 4 bytes |
| 08:04 |
VanessaE |
nore: how so? if you have that "is exposed to sunlight" bit set, then you've got 7 bits of actual light range, and you know you have to include the sunlight into the calc, but you still make it fall off short enough that that never happens |
| 08:04 |
nore |
gregorycu: we should try it |
| 08:05 |
VanessaE |
and afaik sunlight is visible everywhere except in a cave with no direct shaft to the surface, right? |
| 08:05 |
gregorycu |
Well, in the meantime, i'm fixing a memory leak bug :P |
| 08:05 |
VanessaE |
(well a cave or any enclosed space anyways) |
| 08:05 |
nore |
VanessaE: quite so |
| 08:05 |
nore |
there is another problem if we decide to change the lighting format though |
| 08:06 |
nore |
that means old maps have to be converted |
| 08:06 |
VanessaE |
ok, so if a cave is more than 127 nodes deep (let's assume sunlight falls off at one step per node for simplicity), wouldn't the cave be just plain black by then anyway? |
| 08:06 |
VanessaE |
isn't there a map format flag in the file? |
| 08:06 |
nore |
and this is a non-trivial operation |
| 08:07 |
nore |
since it requires reloading a whole map column to fix the lighting in it :/ |
| 08:07 |
VanessaE |
well, I'm suggesting by that, not to convert the map at all. |
| 08:07 |
VanessaE |
if it's old-format, just leave it that way and let the existing methods handle it. |
| 08:07 |
nore |
hm... |
| 08:07 |
VanessaE |
or convert it a block at a time, when blocks are *saved* |
| 08:08 |
VanessaE |
(I assume there's at least a header per mapblock) |
| 08:08 |
nore |
VanessaE: the problem is that you have to convert blocks above before converting a block |
| 08:08 |
VanessaE |
hm |
| 08:09 |
nore |
hm, actually we could do even better |
| 08:09 |
VanessaE |
? |
| 08:09 |
gregorycu |
When loading a block, if it's in the legacy format, you just have to upscale the values |
| 08:10 |
nore |
in a mapblock, use param1 only for artificial light |
| 08:10 |
gregorycu |
Normalise during loading and call it a day |
| 08:10 |
nore |
and map a *2d* array to say how deep sunlight goes in every column |
| 08:10 |
VanessaE |
nore: that's what basically I'm suggesting. |
| 08:10 |
nore |
gregorycu: the problem is sunlight vs. not sunlight actually |
| 08:11 |
kahrl |
what do you do about the pitch black shadows you'll get under every tree? |
| 08:11 |
VanessaE |
compute the REAL light based on max(stored, computed_sunlight_at_pos) |
| 08:11 |
nore |
kahrl: ah, that was why I feared the sharp sunlight cutoff |
| 08:11 |
nore |
but couldn't find an example :/ |
| 08:12 |
VanessaE |
hrm. |
| 08:13 |
|
Krock joined #minetest-dev |
| 08:13 |
gregorycu |
Well, there should be ambient light outside |
| 08:13 |
gregorycu |
So you shouldn't get those perfectly black nodes |
| 08:14 |
nore |
gregorycu: how do you define "outside" then? |
| 08:14 |
gregorycu |
You'd use the 2d array you suggested |
| 08:14 |
VanessaE |
anywhere that isn't cave air and isn't lit by an artificial source? |
| 08:15 |
gregorycu |
The further down you get, less ambient light |
| 08:15 |
nore |
well, under a tree you would be inside because there would be leaves above you... |
| 08:21 |
|
loggingbot_ joined #minetest-dev |
| 08:21 |
|
Topic for #minetest-dev is now Minetest core development and maintenance. Freeze at 2016-03-23 (2 weeks). Last release: 0.4.13, Aug 20 2015. Chit-chat goes to #minetest. Consider this instead of /msg celeron55. http://irc.minetest.ru/minetest-dev/ http://dev.minetest.net/ |
| 08:23 |
gregorycu |
What's the proper way to get this: https://github.com/freeminer/freeminer/commit/551eb8515450be4167504f12653d111e09b0ee32 into a PR? |
| 08:24 |
gregorycu |
I was going to create a branch for this, and somehow cherry-pick the commit over |
| 08:24 |
VanessaE |
add ".patch" to the URL and save-page-as... |
| 08:24 |
VanessaE |
then apply, commit, push, and make a PR |
| 08:24 |
nore |
gregorycu: if I understood correctly, you made your own change without looking at the code, right? |
| 08:24 |
gregorycu |
Yes |
| 08:24 |
nore |
use your own then, it would be better |
| 08:25 |
nore |
because freeminer is GPLv3, so incompatible :/ |
| 08:25 |
gregorycu |
oh |
| 08:25 |
gregorycu |
I wasn't aware of that |
| 08:25 |
gregorycu |
What are we? |
| 08:25 |
nore |
LGPL |
| 08:25 |
VanessaE |
LGPL. |
| 08:25 |
gregorycu |
Nice |
| 08:25 |
nore |
(v2.1 IIRC) |
| 08:26 |
gregorycu |
Ok, making a PR with my own code then |
| 08:26 |
gregorycu |
(this is what I get for trying to do the right thing lol) |
| 08:26 |
gregorycu |
Thanks V for the help too |
| 08:26 |
VanessaE |
sure. |
| 08:29 |
nrzkt |
nore, freeminer is GPLv3 but this source code is in the MT source tree |
| 08:29 |
nrzkt |
it's not original FreeMiner code |
| 08:30 |
nrzkt |
also this fix is trivial and could be ported easily |
| 08:32 |
gregorycu |
So... |
| 08:32 |
gregorycu |
Do what V said? |
| 08:32 |
nrzkt |
V ? |
| 08:32 |
gregorycu |
VanessaE |
| 08:33 |
VanessaE |
nrzkt: does the whole sweat-of-the-brow argument actually work in this context? |
| 08:33 |
nrzkt |
yes |
| 08:33 |
nrzkt |
it's the original MT code |
| 08:33 |
nrzkt |
this source for example cannot be imported into MT https://github.com/freeminer/freeminer/blob/master/src/circuit.cpp |
| 08:34 |
nrzkt |
because it's an original FreeMiner code |
| 08:34 |
nrzkt |
in fact FreeMiner cannot change the original MT licence without all its authors agreement |
| 08:35 |
kahrl |
the LGPL contains a clause that allows anyone to relicense a LGPL work under GPL |
| 08:35 |
|
AnotherBrick joined #minetest-dev |
| 08:35 |
nrzkt |
kahrl, without authors agreement ? |
| 08:35 |
kahrl |
yes |
| 08:35 |
nrzkt |
fucking license i didn't know that |
| 08:35 |
kahrl |
well, the author agreed to it already by licensing under the LGPL |
| 08:36 |
nrzkt |
but a so trivial fix cannot be a license patent problem |
| 08:36 |
nrzkt |
port it |
| 08:36 |
nrzkt |
it's not a special freeminer engineering it's just a trivial fix |
| 08:36 |
kahrl |
dunno, IANAL |
| 08:36 |
kahrl |
and it probably depends on the jurisdiction |
| 08:37 |
nrzkt |
i don't think it's in freeminer interest to do a patent for 2 lines of code fix :p |
| 08:37 |
VanessaE |
that's already been tested in court anyways |
| 08:37 |
nrzkt |
if you want to be the best, ask proller if you can backport fix to upstream |
| 08:40 |
gregorycu |
lol |
| 08:40 |
nore |
well, gregorycu has made another fix anyway, so it's not a problem |
| 08:40 |
nore |
we don't need to use freeminer code :) |
| 08:40 |
VanessaE |
nice timing, proller :) |
| 08:41 |
nore |
sofar: paramat: see my comment on game#862 |
| 08:41 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/862 -- Tnt: Improvements by sofar |
| 08:41 |
gregorycu |
I generally try to reduce changes between major projects |
| 08:41 |
gregorycu |
If it's bugfixes |
| 08:41 |
|
proller joined #minetest-dev |
| 08:41 |
gregorycu |
By changes, I mean differences |
| 08:47 |
VanessaE |
well have fun blowing stuff up. I'm off. |
| 09:06 |
|
PilzAdam joined #minetest-dev |
| 09:27 |
|
paramat joined #minetest-dev |
| 09:28 |
gregorycu |
Indeed |
| 09:28 |
paramat |
gregorycu i think celeron55 might consider making you a core dev if you are happy to |
| 09:29 |
gregorycu |
We've already discussed it, that's cool |
| 09:29 |
paramat |
ok |
| 09:29 |
gregorycu |
In the short term, I am not |
| 09:29 |
gregorycu |
https://github.com/minetest/minetest/issues/3566 |
| 09:29 |
gregorycu |
While I have you |
| 09:29 |
gregorycu |
I made a change that ended up being very similar to the patch mentioned here |
| 09:29 |
gregorycu |
Which you say didn't fix the issue |
| 09:30 |
gregorycu |
"Tested that commit and unfortunately doesn't fix this." |
| 09:30 |
gregorycu |
What is "this" |
| 09:30 |
gregorycu |
You specifically mean the memory leak? |
| 09:30 |
paramat |
correct |
| 09:30 |
gregorycu |
Actually, sorry. The lighting is mentioned latter in the chain. |
| 09:31 |
gregorycu |
Ok |
| 09:33 |
gregorycu |
I'm reasonably confident that that commit and commits like it would fix the memory leak |
| 09:34 |
paramat |
hm it didn't seem to for me, but i hope i'm wrong. maybe it would fix other problems |
| 09:35 |
gregorycu |
If I were to author a PR, would you mind testing it the same way? |
| 09:35 |
|
Obani joined #minetest-dev |
| 09:36 |
paramat |
go ahead i'll test it |
| 09:36 |
paramat |
the issue bothers me |
| 09:37 |
paramat |
i also need to reproduce in my own mod code, in case there's something about the roomgen mod |
| 09:57 |
|
Calinou joined #minetest-dev |
| 10:12 |
|
nyje joined #minetest-dev |
| 10:29 |
paramat |
nore sfan5 PilzAdam game#980 |
| 10:29 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/980 -- Start timer as well when moving items around. by sofar |
| 10:31 |
|
Fixer joined #minetest-dev |
| 10:43 |
paramat |
nore sfan5 PilzAdam i feel this should be refused and closed, opinions? game#726 |
| 10:43 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/726 -- Split default into smaller mods |
| 10:43 |
Obani |
paramat, -1 for me :p |
| 10:53 |
Krock |
refused and closed. would break many mods |
| 10:53 |
Krock |
err actually no |
| 10:56 |
paramat |
#1442 is a possible close |
| 10:56 |
ShadowBot |
https://github.com/minetest/minetest/issues/1442 -- Colliding with unloaded blocks at high velocities is problematic |
| 11:01 |
paramat |
closed #1449 and added it into 3615 |
| 11:01 |
ShadowBot |
https://github.com/minetest/minetest/issues/1449 -- No tool wear out when hitting entities. |
| 11:02 |
paramat |
closed #1452 |
| 11:02 |
ShadowBot |
https://github.com/minetest/minetest/issues/1452 -- Inventory images are flipped upside down - Android Nexus 10 |
| 11:09 |
paramat |
closed #1465 |
| 11:09 |
ShadowBot |
https://github.com/minetest/minetest/issues/1465 -- Hand on right side -only is discriminatory |
| 11:14 |
paramat |
closed #1501 |
| 11:14 |
ShadowBot |
https://github.com/minetest/minetest/issues/1501 -- Shaded block sides should not be blue tinted. |
| 11:14 |
Krock |
\o/ |
| 11:22 |
paramat |
heh. closed and moved #1513 into 1627 |
| 11:22 |
ShadowBot |
https://github.com/minetest/minetest/issues/1513 -- Resize ingame names, depending on distance |
| 11:24 |
paramat |
(re opened 1627) |
| 11:25 |
|
iangp joined #minetest-dev |
| 11:30 |
paramat |
closed and folded #719 into 1533 |
| 11:30 |
ShadowBot |
https://github.com/minetest/minetest/issues/719 -- minetestmapper.py fails to load a world |
| 11:32 |
paramat |
closed #1558 as is fixed |
| 11:32 |
ShadowBot |
https://github.com/minetest/minetest/issues/1558 -- Water pools up at chunk borders |
| 11:37 |
gregorycu |
paramat: I was wrong. In release builds the issue persists |
| 11:37 |
Fixer |
hi gregorycu |
| 11:37 |
paramat |
ok |
| 11:37 |
gregorycu |
Hi there |
| 11:37 |
gregorycu |
How ya been? |
| 11:37 |
Fixer |
gregorycu, any help for #3770 pleaase ? |
| 11:37 |
ShadowBot |
https://github.com/minetest/minetest/issues/3770 -- Fix superflous shader setting updates by ShadowNinja |
| 11:38 |
gregorycu |
Ahh yes, that |
| 11:38 |
Fixer |
gregorycu, i did test it and find lighting bugs |
| 11:38 |
Fixer |
found* |
| 11:38 |
gregorycu |
So I saw :) |
| 11:38 |
gregorycu |
Thanks for that |
| 11:38 |
gregorycu |
More of my stuff gets checked in when people don't find the bugs first |
| 11:39 |
Fixer |
gregorycu, must be 1 or 2 bugs, other than that, it is very promising PR |
| 11:39 |
gregorycu |
Yeah, not everybody gets the same speed improvement |
| 11:39 |
gregorycu |
I think it's probably ATI card on windows, related |
| 11:39 |
Fixer |
gregorycu, gives me very big fps boost, if you can, please look at it |
| 11:39 |
Fixer |
yeah |
| 11:39 |
gregorycu |
I will, I'm checking out a memory leak at the moment |
| 11:40 |
paramat |
Calinou please could you explain what you mean in #1589 ? |
| 11:40 |
ShadowBot |
https://github.com/minetest/minetest/issues/1589 -- New plantlike drawtype causes bad-looking symmetry |
| 11:40 |
Fixer |
gregorycu, thanks |
| 11:41 |
Fixer |
gregorycu, it basically makes performance with shaders like performance without them, but shaders are still working :) |
| 11:42 |
gregorycu |
Yeah, I did profiling on my machine, and saw that most of the time was spent sending information to the shaders, even if the information had not changed |
| 11:42 |
Calinou |
paramat, look at those screenshots, the grass is completely symmetric: https://lut.im/1GybF8HmFB/ZPbGqgdaPnUUN6ww.png | https://lut.im/RZbYPtHOd6/yB4enwXYf3yrOdoN.png |
| 11:42 |
Fixer |
and that kind of performance is fuuuun, game changer for me |
| 11:42 |
Calinou |
it looks awkward |
| 11:42 |
Calinou |
like a human face that would be entirely symmetric |
| 11:43 |
paramat |
ah ok |
| 11:43 |
Fixer |
it is symmetric |
| 11:43 |
paramat |
we would need 2 textures per plant though, don't think it's worth it |
| 11:44 |
Calinou |
not necessarily |
| 11:44 |
gregorycu |
Fixer: I'm glad you like it, hopefully i can fix it very soon |
| 11:44 |
Calinou |
just don't make them symmetric |
| 11:44 |
Calinou |
Minecraft does it, with only 1 texture |
| 11:44 |
Calinou |
and we used to do before 2012 |
| 11:44 |
paramat |
looks ok to me in a voxel world |
| 11:44 |
Calinou |
it might be very bad for HD texture packs... |
| 11:44 |
Fixer |
gregorycu, yes, in almost all scenes there are big boost, in one huge forest with 240 range it gave me 17 to 70 fps or so |
| 11:44 |
Calinou |
a symmetric apple from the Misa pack means it "shines" twice when viewed from some angles |
| 11:44 |
|
sloantothebone joined #minetest-dev |
| 11:45 |
Calinou |
totally unrealistic :P |
| 11:45 |
gregorycu |
Fixer: Wow, that's a lot |
| 11:45 |
Fixer |
yes, i was surprised |
| 11:45 |
paramat |
well, minetest is unrealistic :) |
| 11:45 |
Fixer |
gregorycu, ATI cards are awkward, even guys with different ATI card has big VBO boost, while i have small o_0 |
| 11:46 |
gregorycu |
I haven't even seen VBO yet |
| 11:46 |
Fixer |
gregorycu, it is merged in master |
| 11:46 |
Calinou |
I use a recent build with VBO on my laptop |
| 11:46 |
Fixer |
gregorycu, enabled by default |
| 11:46 |
Calinou |
I still don't always have 60 FPS :( |
| 11:47 |
gregorycu |
Calinou: Just wait for my changes |
| 11:47 |
Calinou |
(i7-4720HQ, 8 GB RAM, GTX 960M) |
| 11:47 |
gregorycu |
Maybe I can give you a couple of extra fps |
| 11:47 |
gregorycu |
Or maybe none |
| 11:47 |
gregorycu |
I'm using a mod where I'm underground |
| 11:47 |
gregorycu |
To test this memory leak |
| 11:48 |
gregorycu |
Gotta figure out how to track it |
| 11:50 |
paramat |
using 1 texture for plantlike will always create symmetry from some directions, unavoidable |
| 11:50 |
|
dfelinto joined #minetest-dev |
| 11:51 |
gregorycu |
Because the back face is a mirror of the front face |
| 11:51 |
paramat |
3770 is not milestone though, so no rush for release |
| 11:51 |
gregorycu |
But it's the best PR yet? |
| 11:52 |
gregorycu |
:) |
| 11:52 |
paramat |
it would be good to have in release, we're just understaffed and est is concentrating on 3848 |
| 11:52 |
|
Obani joined #minetest-dev |
| 11:52 |
gregorycu |
All good, hopefully I can help with some of the stuff |
| 11:53 |
paramat |
also #3894 is higher priority |
| 11:53 |
ShadowBot |
https://github.com/minetest/minetest/issues/3894 -- Add serverside noclip enforcement by est31 |
| 11:53 |
paramat |
yeah it's so good to see you |
| 11:53 |
gregorycu |
I'm no longer team leader at work, so I can spend some time |
| 11:55 |
Calinou |
paramat, no, you can use 1 texture and 4 planes instead of just 2 |
| 11:55 |
Calinou |
4 planes with backfacde c |
| 11:55 |
Calinou |
4 planes with backface culling, instead of 2 without |
| 11:59 |
Fixer |
yes, 3894 will be very good too, I've played sibe by side with cheaters and observed them, noclip prevention is super needed |
| 11:59 |
Fixer |
noclip intrusion + kill or noclip mining for mese and other |
| 12:00 |
Fixer |
slow fly is not mitigated though |
| 12:00 |
Fixer |
but there are some mods |
| 12:01 |
Calinou |
what about increasing dedicated_server_step rate to 20 FPS, to make noclip detection more reliable and with less false positives? |
| 12:01 |
Calinou |
dedicated_server_step = 0.05 |
| 12:01 |
Calinou |
I've noticed only minimal CPU increase |
| 12:01 |
Calinou |
default is 0.1 (10 FPS) |
| 12:02 |
paramat |
ah i see |
| 12:02 |
paramat |
complex though |
| 12:09 |
|
gregorycu joined #minetest-dev |
| 12:10 |
paramat |
closed #1608 we have that feature now |
| 12:10 |
ShadowBot |
https://github.com/minetest/minetest/issues/1608 -- Allow limiting mapgen per world |
| 12:10 |
Calinou |
only 623 issues left! :p |
| 12:11 |
Calinou |
https://github.com/minetest/minetest/issues/34 |
| 12:11 |
Calinou |
you can remove "Feature request:" from the title |
| 12:11 |
Calinou |
since we have a label for it |
| 12:13 |
paramat |
heh. i can't be bothered to edit titles, too busy |
| 12:19 |
celeron55 |
i don't think we have that feature |
| 12:19 |
celeron55 |
isn't 2908 a global setting? |
| 12:19 |
celeron55 |
it's not per-world |
| 12:20 |
celeron55 |
can someone confirm because clearly if this is the case the PR should stay open |
| 12:20 |
celeron55 |
i mean issue |
| 12:21 |
paramat |
yes 2908 is global |
| 12:21 |
paramat |
i'll re open |
| 12:22 |
paramat |
ah i didn't see the 'per world' bit =/ |
| 12:23 |
paramat |
it should really be per-world |
| 12:24 |
paramat |
closed #1710 |
| 12:24 |
ShadowBot |
https://github.com/minetest/minetest/issues/1710 -- Subsystem maintainers |
| 12:25 |
celeron55 |
a lot of things should be per-world really |
| 12:25 |
paramat |
i need to learn how to add a per-world mapgen parameter so if i can i'll do this |
| 12:28 |
Calinou |
paramat, https://github.com/minetest/minetest/issues/1670 -> I don't understand, what do you mean by node defs? |
| 12:28 |
Calinou |
I'm pretty sure glow mapping can do things node defs can't |
| 12:29 |
paramat |
erm |
| 12:30 |
paramat |
is it not simpler to just define the particular node as glowing? instead of detecting a special texture |
| 12:30 |
Calinou |
https://lut.im/zhrIIN2MAG/PKza11QUh6ObNpZa.png |
| 12:31 |
paramat |
oh 'regardless of light level' |
| 12:31 |
Calinou |
the green "squares" are glow mapped here |
| 12:31 |
Calinou |
they do not emit light by themselves (there's a light entity nearby though, to emulate that) |
| 12:31 |
Calinou |
but they'll glow even in the darkness. |
| 12:31 |
Calinou |
and other areas of the texture can stay dark |
| 12:31 |
Calinou |
this is exactly the point of glow mapping |
| 12:31 |
Calinou |
even Quake 3 did it :) |
| 12:31 |
paramat |
sorry i didn't understand |
| 12:32 |
paramat |
i thought you played quake, as you like fast running |
| 12:32 |
paramat |
:) |
| 12:33 |
Calinou |
https://lut.im/CsYgKpBwAN/I5U8JPMUoCgtw1fM.png |
| 12:33 |
Calinou |
here, I removed all lights |
| 12:34 |
paramat |
i see |
| 12:36 |
paramat |
closed and folded #1693 into 2731 as that has more discussion |
| 12:36 |
ShadowBot |
https://github.com/minetest/minetest/issues/1693 -- Stuff about keybindings |
| 12:53 |
paramat |
closed #1815 |
| 12:53 |
ShadowBot |
https://github.com/minetest/minetest/issues/1815 -- Incomplete mapblocks being saved to database |
| 12:56 |
|
STHGOM joined #minetest-dev |
| 13:00 |
Obani |
We would need Github moderators o.o |
| 13:03 |
paramat |
enough for now, 20 closed :) |
| 13:21 |
gregorycu |
What did you close? |
| 13:24 |
|
iangp joined #minetest-dev |
| 13:33 |
paramat |
those i linked in this channel |
| 13:36 |
gregorycu |
Ahh issues |
| 13:36 |
gregorycu |
Nice work |
| 13:36 |
gregorycu |
Maybe do another 20? |
| 13:39 |
paramat |
heh hopefully |
| 13:42 |
|
nrzkt joined #minetest-dev |
| 13:44 |
sfan5 |
paramat: +1 for game#980 |
| 13:44 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/980 -- Start timer as well when moving items around. by sofar |
| 13:44 |
sfan5 |
also closed the 'seperate into mods' issue |
| 13:45 |
PilzAdam |
paramat, #980 is good, but you should really reconsider your view on #978 |
| 13:46 |
ShadowBot |
https://github.com/minetest/minetest/issues/980 -- Allow make fast bots for servers load testing by proller |
| 13:46 |
ShadowBot |
https://github.com/minetest/minetest/issues/978 -- try catch missing in lua called c++ errors |
| 13:46 |
PilzAdam |
* game |
| 13:48 |
paramat |
sure, i guess i mean sometimes it has to be done |
| 13:49 |
paramat |
but of course best avoided if possible |
| 13:55 |
|
damiel joined #minetest-dev |
| 13:57 |
|
paramat left #minetest-dev |
| 14:01 |
|
nrzkt joined #minetest-dev |
| 14:04 |
gregorycu |
Sweet jesus |
| 14:04 |
gregorycu |
*(void **)(lua_newuserdata(L, sizeof(void *))) = o; |
| 14:05 |
gregorycu |
I'm a pretty damn good programmer and I'm not sure I want to try to make sense of that |
| 14:05 |
sfan5 |
huh |
| 14:05 |
sfan5 |
casts the result of lua_newuserdata(...) to a void** |
| 14:05 |
sfan5 |
dereferences that pointer |
| 14:06 |
sfan5 |
and assign o to it |
| 14:07 |
gregorycu |
You sound like you know what you're talking about then |
| 14:07 |
|
kaadmy joined #minetest-dev |
| 14:09 |
gregorycu |
Hang on, I'll try to get my next question answered by the debugger |
| 14:14 |
gregorycu |
Who is our Lua expert? |
| 14:14 |
gregorycu |
In terms of the C API? |
| 14:21 |
|
Darcidride joined #minetest-dev |
| 14:38 |
PilzAdam |
gregorycu, you are |
| 14:38 |
gregorycu |
Curses |
| 14:38 |
PilzAdam |
:-p |
| 14:46 |
gregorycu |
I don't even like Lua |
| 14:46 |
gregorycu |
This doesn't seem fair |
| 15:01 |
|
hmmmm joined #minetest-dev |
| 15:14 |
|
poignardazur_ joined #minetest-dev |
| 15:27 |
|
proller joined #minetest-dev |
| 15:35 |
|
Samson1 joined #minetest-dev |
| 15:55 |
|
Darcidride joined #minetest-dev |
| 16:00 |
|
paramat joined #minetest-dev |
| 16:01 |
paramat |
gregorycu ShadowNinja is good with API |
| 16:01 |
gregorycu |
Thanks |
| 16:01 |
gregorycu |
Any reason why est31 isn't on IRC? |
| 16:02 |
paramat |
he says he concentrates better away from IRC, also busy with life stuff |
| 16:13 |
gregorycu |
minetest memory churn is pretty bad |
| 16:15 |
gregorycu |
I told windows to reclaim as much memory as it could from minetest. It reclaimed 1.1GB out of 1.3GB |
| 16:19 |
paramat |
! |
| 16:20 |
gregorycu |
Actually was closer to 1.2GB |
| 16:20 |
gregorycu |
It went from using 1.3GB to 140MB |
| 16:21 |
|
Krock joined #minetest-dev |
| 16:31 |
|
Player_2 joined #minetest-dev |
| 16:37 |
gregorycu |
paramat: It's pitch black at night |
| 16:38 |
gregorycu |
Standing next to a torch |
| 16:38 |
paramat |
hm |
| 16:38 |
gregorycu |
Actually, you can see the torches, but they emit no light |
| 16:38 |
gregorycu |
I have a screenie |
| 16:39 |
paramat |
so, is it possible that this mod's rapid adding of 700 schematics per mapchunk is what is exposing the problem? |
| 16:40 |
gregorycu |
I think this is a different issue where the lighting isn't calculated for night for placed schematics |
| 16:40 |
paramat |
ok |
| 16:41 |
paramat |
by 'exposing/triggering the problem' i mean the memory leak |
| 16:42 |
gregorycu |
Everytime you place a schematic it leaks |
| 16:42 |
gregorycu |
Do a lot of them, and you'd notice :) |
| 16:45 |
gregorycu |
http://i.imgur.com/SNkHfZJ.jpg |
| 16:45 |
gregorycu |
That's with noclip |
| 16:46 |
gregorycu |
Off to bed for me |
| 16:46 |
|
rubenwardy joined #minetest-dev |
| 16:50 |
|
rubenwardy joined #minetest-dev |
| 17:51 |
|
rubenwardy joined #minetest-dev |
| 18:25 |
|
STHGOM joined #minetest-dev |
| 18:39 |
|
rubenwardy joined #minetest-dev |
| 19:06 |
|
davisonio joined #minetest-dev |
| 19:41 |
|
Gael-de-Sailly joined #minetest-dev |
| 19:56 |
paramat |
hmmmm if possible could you have a quick look at #3927 sometime? |
| 19:56 |
ShadowBot |
https://github.com/minetest/minetest/issues/3927 -- Decoration API: Allow rotation of simple decorations by paramat |
| 19:57 |
sofar |
p2-based rotation? |
| 19:58 |
paramat |
yeah :) |
| 19:58 |
sofar |
can I request a more generic method instead? |
| 19:58 |
sofar |
can I request that the interface allows passing *any* p2 value? |
| 19:59 |
paramat |
i thought you would heh |
| 19:59 |
sofar |
oh, wait, one sec |
| 19:59 |
* sofar |
reads |
| 20:01 |
* sofar |
gets irq'd |
| 20:03 |
sofar |
paramat: why would you _not_ use param2 directly here? |
| 20:04 |
sofar |
I can see "rotation" sounds simpler, but what if someone wants to put up some stack of logs with param2 >= 4 ? |
| 20:05 |
paramat |
oh good point |
| 20:05 |
paramat |
i could support all 24 values |
| 20:05 |
paramat |
ok thanks |
| 20:06 |
sofar |
yes, and therefore it could be used to place wild plants with my various plant mesh changes |
| 20:08 |
paramat |
although when someone wants random rotation they may want NESW or all 24 |
| 20:08 |
paramat |
both options may be needed |
| 20:10 |
paramat |
i might add some flags to enable this |
| 20:11 |
paramat |
sorry hmmmmm to disturb you, it's WIP now |
| 20:13 |
sofar |
no problem, you know I'm a fan of having param2 modifyable through decorations for obvious reasons :) |
| 20:15 |
|
blaze joined #minetest-dev |
| 20:19 |
nore |
sofar: about game#862 |
| 20:19 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/862 -- Tnt: Improvements by sofar |
| 20:19 |
sofar |
yah |
| 20:19 |
nore |
just two stacks of ~4 TNTs were enough |
| 20:19 |
nore |
one, then the other |
| 20:19 |
nore |
but they were underwater |
| 20:19 |
sofar |
oh |
| 20:20 |
sofar |
on the bottom of a lake or something? |
| 20:20 |
nore |
I don't know if that may have caused it or not |
| 20:20 |
nore |
hm, yeah |
| 20:20 |
sofar |
weird, I know I tried underwater plenty |
| 20:20 |
sofar |
did any items drop? |
| 20:21 |
nore |
let me try again |
| 20:21 |
sofar |
oh, server hung. did you restart and take a look? |
| 20:21 |
nore |
I couldn't stop it without C-c |
| 20:21 |
nore |
so there was just one tnt less |
| 20:21 |
sofar |
yeah, but when you restart the server the dropped items should be there |
| 20:22 |
sofar |
would be interesting to know if there were items after restarting or not |
| 20:22 |
sofar |
since that may help to narrow it down |
| 20:22 |
nore |
wait, that may be related |
| 20:22 |
nore |
there is sand at the bottom |
| 20:23 |
sofar |
I have seen a few warnings when you explode 100+ TNT. but that's to be expected as you're adding too many items :) |
| 20:23 |
sofar |
besides that didn't crash my server |
| 20:23 |
sofar |
falling sand? |
| 20:23 |
nore |
just triggered it again |
| 20:23 |
sofar |
yeah, I wonder if that's the problem, falling nodes |
| 20:23 |
nore |
sand gets ejected in all directions |
| 20:23 |
nore |
but there is no falling sand |
| 20:24 |
sofar |
the little sand item cubes or whole falling sand nodes? |
| 20:26 |
|
newuser_ joined #minetest-dev |
| 20:26 |
nore |
items |
| 20:27 |
sofar |
that's fine, that's one of the patches in the series |
| 20:28 |
nore |
there, server hung up again :/ |
| 20:28 |
nore |
and this time, I'm not getting any error messages |
| 20:29 |
nore |
but mt is at 180% cpu |
| 20:29 |
paramat |
related to the falling.lua fix? #3925 |
| 20:29 |
ShadowBot |
https://github.com/minetest/minetest/issues/3925 -- stop falling.lua error by tenplus1 |
| 20:30 |
paramat |
or maybe Rui's recent changes to falling |
| 20:30 |
nore |
paramat: I don't think it's falling... |
| 20:30 |
sofar |
nore's errors are different, though |
| 20:30 |
nore |
unrelated to 3925 too |
| 20:31 |
sofar |
nore: try in underwater dirt? |
| 20:31 |
sofar |
to eliminate falling nodes? |
| 20:31 |
sofar |
could even be water |
| 20:32 |
paramat |
ok |
| 20:35 |
nore |
sofar: I could reproduce it without sand |
| 20:35 |
nore |
trying without water now |
| 20:35 |
sofar |
could, or can? |
| 20:35 |
sofar |
english subtlety, but important nonetheless |
| 20:35 |
nore |
I just reproduced it |
| 20:36 |
sofar |
so "can" :) |
| 20:36 |
nore |
sorry for that english bug :) |
| 20:36 |
sofar |
de rien ;) |
| 20:38 |
nore |
and... again, without any water this time |
| 20:38 |
nore |
3 stacks of 4 tnts buried in the ground |
| 20:38 |
nore |
one after another, in separate locations |
| 20:41 |
nore |
oh, and there is another problem: if server shutdowns between tnt is triggered and explodes, it never explodes it looks like |
| 20:41 |
nore |
same for chain reactions |
| 20:42 |
nore |
I'm wondering if it might be related to the huge number of particles it generates |
| 20:42 |
nore |
sofar: were you able to reproduce the bug? |
| 20:44 |
|
davisonio joined #minetest-dev |
| 20:48 |
|
STHGOM joined #minetest-dev |
| 20:51 |
|
est31 joined #minetest-dev |
| 21:14 |
sofar |
nore: I'm at work, can't run MT here anyway |
| 21:16 |
|
DFeniks joined #minetest-dev |
| 21:19 |
|
paramat left #minetest-dev |
| 22:03 |
|
rubenwardy joined #minetest-dev |
| 22:17 |
|
rubenwardy_ joined #minetest-dev |
| 22:28 |
|
iangp joined #minetest-dev |
| 22:34 |
|
Player_2 joined #minetest-dev |
| 22:42 |
rubenwardy |
0.4.13-stable can't connect to 0.4.13-dev, it seems. |
| 22:47 |
ssieb |
rubenwardy: is there an error message? |
| 22:48 |
rubenwardy |
says "serialisation error" and something about the wrong nodebox version |
| 22:49 |
rubenwardy |
it was 0.4.13-latest-dev for server and 0.4.13-early-dev from a month or so ago |
| 22:52 |
rubenwardy |
nevermind, it appears it was something to do with the version I was running |
| 22:52 |
rubenwardy |
I think I was running a vbo or a farmesh build |
| 22:54 |
sofar |
old client, new server |
| 22:54 |
rubenwardy |
It works with 0.4.13-stable, just not with the patch I was running. At least I can't reproduce it |
| 23:13 |
Fixer |
hold on |
| 23:14 |
Fixer |
rubenwardy, unsupported NodeBox version |
| 23:15 |
Fixer |
rubenwardy, was reported here too https://forum.minetest.net/viewtopic.php?f=10&t=8494&start=1500#p210464 |
| 23:31 |
|
Dragonop joined #minetest-dev |
