Time |
Nick |
Message |
00:11 |
|
OldCoder joined #minetest-dev |
00:32 |
|
asl97 joined #minetest-dev |
00:56 |
|
Lunatrius joined #minetest-dev |
01:28 |
|
sloantothebone joined #minetest-dev |
01:29 |
|
sloantothebone joined #minetest-dev |
01:30 |
|
sloantothebone joined #minetest-dev |
01:32 |
|
sloantothebone joined #minetest-dev |
01:33 |
|
sloantothebone joined #minetest-dev |
01:57 |
|
Etzos joined #minetest-dev |
02:11 |
est31 |
https://sourceforge.net/p/irrlicht/bugs/435 lets see what the irrlicht crowd sais |
02:12 |
est31 |
if they fix it, we copy the fixed function into our codebase |
02:14 |
|
RealBadAngel joined #minetest-dev |
02:21 |
|
diemartin joined #minetest-dev |
02:45 |
|
sloantothebone joined #minetest-dev |
02:46 |
|
sloantothebone joined #minetest-dev |
02:47 |
|
turtlejerky joined #minetest-dev |
02:47 |
|
sloantothebone joined #minetest-dev |
02:49 |
est31 |
hmmmm, how can I create a singlenode water mapgen? |
02:49 |
|
sloantothebone joined #minetest-dev |
02:49 |
est31 |
I want to use valgrind to check for the leak |
02:49 |
est31 |
but mapgen is too slow |
02:49 |
est31 |
oh wait |
02:50 |
|
sloantothebone joined #minetest-dev |
02:52 |
|
sloantothebone joined #minetest-dev |
03:01 |
sloantothebone |
Hi |
03:02 |
sloantothebone |
What would I do if I wanted to add a hundred different types of nodes without defining every single one by hand? Forgive me i this question belongs in #minetest |
03:07 |
est31 |
sloantothebone, make a loop |
03:07 |
est31 |
also yes it belongs in #minetest |
03:09 |
|
Higgs_Bonbon joined #minetest-dev |
03:36 |
turtlejerky |
what exactly does adding "--depth 1" when git cloning minetest do? |
03:37 |
asl97 |
iirc, make it clone only the head commit |
03:38 |
Wayward_One |
interesting... further testing is to come, but after redownloading minetest and building the apk, the version tag reads 0.4.11-13e100e1 |
03:40 |
Wayward_One |
And the only thing I can think of that I did different was not use --depth 1 |
03:43 |
|
sloantothebone joined #minetest-dev |
03:47 |
est31 |
if just paramat were here |
03:47 |
est31 |
I've run a server with valgrind |
03:47 |
est31 |
and found out that singlenode and v7 mapgen leak both the same amount |
03:47 |
est31 |
so its either a client leak |
03:48 |
est31 |
or the leak is somewhere else |
03:51 |
|
OldCoder joined #minetest-dev |
04:31 |
|
yang2003 joined #minetest-dev |
04:32 |
sloantothebone |
How do I make a golem mod? |
04:32 |
sloantothebone |
Like is it possible to replace a node structure with an entity |
04:32 |
est31 |
--> #minetest |
04:32 |
sloantothebone |
ok |
04:33 |
asl97 |
on_place, check for the four node below it, remove the node if all is there and add entity |
04:33 |
asl97 |
oh, and remove the head too |
04:36 |
|
yang2003 left #minetest-dev |
05:02 |
hmmmm |
est31, alias default:water_source to mapgen_singlenode |
05:05 |
est31 |
ok |
05:05 |
est31 |
hmmmm, you know more about paramat's leak? |
05:06 |
Wayward_One |
so, after completely deleting minetest a second time, redownloading, and building the apk again, the version tag still reads 0.4.11-13e100e1 |
05:06 |
hmmmm |
i don't know anything about paramat's leak |
05:06 |
est31 |
k |
05:06 |
hmmmm |
i suggested he try looking at the noise functions since that's the only place where the mapgen allocates memory, and you just said it happens in singlenode where there is literally no noise at all |
05:06 |
hmmmm |
and he mentioned it happens at least as far back as 0.4.9 |
05:06 |
est31 |
I dunno whether there is a leak |
05:07 |
hmmmm |
me neither |
05:07 |
est31 |
I just get the exact same number when I run valgrind |
05:07 |
hmmmm |
there is a way to tell |
05:07 |
asl97 |
Wayward_One: download using git? if so, check your remote and branch |
05:08 |
hmmmm |
i.e. load a crapton of blocks, then look downward until all of the surrounding blocks unloaded due to inactivity timeout, and then compare the memory usage |
05:08 |
est31 |
he sais its only happening when mapgen happens |
05:08 |
hmmmm |
and do this several times |
05:08 |
est31 |
lets first find out whether its on the server, or on the client |
05:09 |
Wayward_One |
asl97: I always do. I've been making regular builds on the forum for a while, all built the same way, and never had this happen before |
05:10 |
asl97 |
maybe someone mess up the tree |
05:10 |
asl97 |
by using force push |
05:10 |
est31 |
there has been a 0.4.11 bug on normal already |
05:11 |
est31 |
perhaps it reappears on android |
05:11 |
Wayward_One |
Could be |
05:11 |
Wayward_One |
This is the first I've seen of it though |
05:11 |
|
blaaaaargh joined #minetest-dev |
05:15 |
est31 |
hmmmm, I think the leak paramat observes is no mapgen leak |
05:15 |
est31 |
I connected to a localhost game, and flew around a bit |
05:15 |
est31 |
and got 200 MB at client, 40-60 MB at server |
05:16 |
est31 |
at the start both processes had the same memory amount |
05:16 |
est31 |
its a client leak |
05:17 |
est31 |
when I disconnect from the server, the client doesnt deallocate those 200MB |
05:18 |
est31 |
perhaps we don't drop the meshes I dunno |
05:21 |
hmmmm |
you used separate processes for server and client? hmm |
05:22 |
est31 |
yea, connected to localhost |
05:22 |
est31 |
one instance with ---server |
05:22 |
est31 |
other from client tab |
05:29 |
hmmmm |
hrum |
05:41 |
est31 |
where was that lua doc website with the beer mug again? |
05:41 |
|
asl joined #minetest-dev |
05:43 |
|
Hunterz joined #minetest-dev |
05:44 |
est31 |
thats the page with the best documentation what happens |
05:44 |
est31 |
any ideas hmmmm ? |
05:48 |
est31 |
ah its http://pgl.yoyo.org/luai/i/about |
05:48 |
est31 |
thanks though |
05:50 |
* VanessaE |
peeks in |
06:02 |
hmmmm |
haa |
06:02 |
hmmmm |
est31: i wouldn't have known anyway |
06:08 |
|
OldCoder joined #minetest-dev |
06:11 |
est31 |
hmmmm, can I push this https://github.com/est31/minetest/commit/3514ac685fab28885f9e47cd9c3bc33f90e64eab |
06:13 |
est31 |
its to make the checkbox work in real time |
06:15 |
hmmmm |
is const std::string value the actual callback prototype? |
06:15 |
est31 |
yes |
06:15 |
hmmmm |
wth |
06:15 |
est31 |
should be changed IMO |
06:15 |
est31 |
can I push another commit that changes it |
06:15 |
hmmmm |
probably a typo |
06:15 |
hmmmm |
we should fix that sometime |
06:15 |
hmmmm |
yes |
06:15 |
est31 |
also its const std::string settingname |
06:15 |
hmmmm |
also what's with the unix_lowercase_underscore_function_style_in_methods |
06:15 |
est31 |
not value |
06:15 |
est31 |
mistake by me |
06:16 |
est31 |
ah you want it readSettings |
06:16 |
hmmmm |
http://dev.minetest.net/Code_style_guidelines#Classes |
06:16 |
est31 |
yea |
06:17 |
est31 |
I keep forgetting it |
06:22 |
|
RealBadAngel joined #minetest-dev |
06:26 |
est31 |
pushing the two commits then : https://github.com/est31/minetest/commits/master |
06:26 |
est31 |
(if my make succeeds) |
06:27 |
hmmmm |
hold on a minute |
06:27 |
hmmmm |
you don't give even 5 minutes for people to look |
06:27 |
hmmmm |
and besides, you haven't gotten a second +1 |
06:27 |
est31 |
<hmmmm> yes |
06:27 |
hmmmm |
so? i agreed to the idea that it should be changed. there's a difference between an idea and its implementation. |
06:28 |
hmmmm |
what if you made a fatal typo or something in that implementation? that's the reason why code reviews exist |
06:28 |
est31 |
ok then, can you look at it |
06:28 |
hmmmm |
yes, i did |
06:28 |
hmmmm |
it's fine |
06:29 |
est31 |
can I push then? or do you believe I need a third (or second or whatever) +1? |
06:29 |
hmmmm |
fine |
06:29 |
hmmmm |
maybe the number of those in agreement should vary based on size/complexity |
06:30 |
RealBadAngel |
anybody remember why BS was introduced? |
06:30 |
est31 |
that would be a good rule change |
06:30 |
hmmmm |
so that people don't get confused between map coordinates and irrlicht coordinates |
06:30 |
est31 |
however question is who determines that number |
06:30 |
RealBadAngel |
can it be that need for camera offset is caused by that? |
06:31 |
hmmmm |
of course not, irrlicht coordinates are floating point numbers |
06:31 |
VanessaE |
RealBadAngel: nah, camera offset (if you mean the every-200-nodes thing) is due to precision errors in irrlicht renbdering |
06:31 |
VanessaE |
rendering* |
06:32 |
VanessaE |
damn it, ninja'd |
06:32 |
VanessaE |
shit used to get really glitchy out past around 1500-2000 nodes |
06:32 |
hmmmm |
minecraft had the same problem |
06:32 |
hmmmm |
heh |
06:32 |
hmmmm |
that added to the mystique of the 'far lands' |
06:34 |
hmmmm |
so does anybody volunteer to work on the remainder of what needs to be done on the minimap? |
06:34 |
VanessaE |
yeah except farlands sometimes looks kinda nice in the pics I've seen, whereas minetest...er... has nothing out there :) |
06:35 |
RealBadAngel |
minimap project is far from being complete, those are not "remainders" |
06:36 |
RealBadAngel |
but im not sure if i know the minimap code good enough :P |
06:36 |
hmmmm |
it's a decent basis |
06:37 |
RealBadAngel |
theres API to be done, players markers, waypoints, creating map textures for use within mods, etc |
06:37 |
hmmmm |
i have two more uses for the minimap: map previewer and an embedded minetestmapper that's officially supported and maintained |
06:37 |
est31 |
"embedded"? |
06:37 |
hmmmm |
yeah |
06:37 |
hmmmm |
you know the functionality of minetestmapper? |
06:38 |
est31 |
yes |
06:38 |
hmmmm |
well I want that in minetest |
06:38 |
est31 |
making png files |
06:38 |
hmmmm |
to generate an image map as map is generated |
06:38 |
hmmmm |
a couple things though |
06:38 |
hmmmm |
it wouldn't be available immediately in minetestserver due to the *heavy* dependencies on irrlicht |
06:38 |
est31 |
minetestmapper has far more maintained forks than minetest |
06:39 |
est31 |
not sure whether to add another one |
06:39 |
hmmmm |
minetestmapper is slow |
06:39 |
est31 |
forks or compliant reimplementations |
06:39 |
hmmmm |
if you map on the go, you could theoretically have a website with a near real-time map of your minetest server |
06:39 |
hmmmm |
now that's sorta cool |
06:40 |
est31 |
a person running a geo mapping tools company has even implemented a live mapper |
06:40 |
est31 |
https://www.youtube.com/watch?v=8e6eCMXguu8 |
06:40 |
est31 |
dunno how he got that html into minetest |
06:41 |
RealBadAngel |
hmmm, one of the things to do is moving image generation back to mapper thread |
06:41 |
RealBadAngel |
without dep on driver |
06:41 |
RealBadAngel |
just direct creating image in memory |
06:42 |
RealBadAngel |
its trivial, we do need just own setPixel stuff |
06:42 |
hmmmm |
generating the IMAGE is one thing |
06:42 |
hmmmm |
creating a TEXTURE from the image is another |
06:42 |
hmmmm |
the latter of the two is not thread-safe |
06:42 |
VanessaE |
bear in mind that a full size detailed map of a server can run 50 MB in size (*after* pngcrush) |
06:42 |
RealBadAngel |
that you can you make in main |
06:43 |
hmmmm |
VanessaE, which is why it'd probably be a good idea to split it up into chunks |
06:43 |
VanessaE |
and that would only be a fraction of the total map space. |
06:43 |
RealBadAngel |
most time consuming is to create an image |
06:43 |
hmmmm |
but otherwise do you think there's value in concept? |
06:43 |
VanessaE |
hmmmm: right. est31 seems like your leaftest thing is apropos here? |
06:43 |
est31 |
or to generate it in chunks the first place |
06:43 |
|
Calinou joined #minetest-dev |
06:43 |
hmmmm |
Leaftest? |
06:43 |
VanessaE |
yes, I could see it being useful as long as the user is very clear on how to see the map |
06:43 |
est31 |
VanessaE, thats his own leaflet implementation |
06:43 |
VanessaE |
https://github.com/est31/leaftest |
06:43 |
RealBadAngel |
you can save minimapmapblocks to the disk |
06:43 |
RealBadAngel |
outside app can just read them |
06:43 |
VanessaE |
est31: yeah I know, but it's similar to what hmmmm is talking about |
06:44 |
hmmmm |
sloppy map? |
06:44 |
est31 |
with many small tiles |
06:44 |
hmmmm |
well |
06:44 |
est31 |
^ =: sloppy |
06:44 |
hmmmm |
if there's no value in this then there's no value in this |
06:44 |
hmmmm |
that makes less work for us |
06:45 |
hmmmm |
i still want to use Mapper for map previews when selecting a seed |
06:45 |
est31 |
preview is cool though |
06:45 |
est31 |
for seeds and for world selection |
06:45 |
est31 |
store the last despawned pos |
06:45 |
hmmmm |
I wanted to add a map preview since I started coding on minetest |
06:46 |
hmmmm |
never did it though because meh |
06:46 |
hmmmm |
other things had higher priority |
06:46 |
hmmmm |
in any case |
06:46 |
hmmmm |
next up is client side scripting |
06:46 |
hmmmm |
i'm officially starting on that now |
06:46 |
Hunterz |
Im trying implement server console using readline alternate interface (callbacks), or will be better wait for keyboard events in the another thread ? |
06:46 |
hmmmm |
also a minimap becomes a hud item that a script will be able to draw |
06:47 |
hmmmm |
so you'd (in theory) be able to draw a minimap on the main menu as well as having multiple minimaps on the screen in game |
06:47 |
|
chchjesus joined #minetest-dev |
06:47 |
est31 |
it should be a hud item as well as a formspec item. |
06:47 |
hmmmm |
that's another thing |
06:47 |
est31 |
perhaps even moveable and so on |
06:47 |
VanessaE |
hmmmm: now THAT would be useful (map preview on create) |
06:47 |
hmmmm |
you should be able to draw hud items on GUIs as well |
06:48 |
est31 |
but yea thats another thing |
06:48 |
hmmmm |
i need to sort out how HUD and GUIs are going to work |
06:48 |
est31 |
perhaps that should even be server generated. |
06:48 |
hmmmm |
and whether they should be separate at all |
06:48 |
RealBadAngel |
mapper should just produce textures for use in mods |
06:48 |
RealBadAngel |
then you could use them anywhere |
06:48 |
hmmmm |
hunterz: a callback in which thread? |
06:49 |
Hunterz |
in the main server thread |
06:49 |
est31 |
the main difference between hud and gui is that once you have a mouse button, once your mouse is fixed |
06:49 |
hmmmm |
hunterz: if I were writing that, I would handle the console interface in a separate thread and then queue commands to be executed in the server thread |
06:49 |
est31 |
but that difference only shows up on non-android devices |
06:49 |
Hunterz |
ok, thanks |
06:50 |
hmmmm |
est31: alright how does this sound |
06:50 |
hmmmm |
you can draw hud items ANYWHERE on the screen at any time |
06:50 |
hmmmm |
i.e. in game, in the main menu, on a formspec |
06:51 |
hmmmm |
well we call it formspec right now but |
06:51 |
hmmmm |
the GUI elements |
06:51 |
hmmmm |
GUI elements, however, can only be present in a GUI window |
06:51 |
hmmmm |
a GUI window being opened will display the mouse and release the lock on mouse position |
06:51 |
hmmmm |
s/mouse/cursor/ |
06:51 |
RealBadAngel |
hmmmm, have you noticed what minimap on screen element code is? |
06:52 |
est31 |
hrmm |
06:52 |
hmmmm |
RealBadAngel: what? |
06:52 |
RealBadAngel |
its modified IGUI element |
06:52 |
est31 |
so one use case of drawing hud elements in gui is minimap |
06:52 |
hmmmm |
what? |
06:52 |
est31 |
any other usecase? |
06:52 |
hmmmm |
I thought it was a mesh |
06:53 |
RealBadAngel |
it uses to display the mesh |
06:53 |
hmmmm |
est31: yea |
06:53 |
hmmmm |
how about this |
06:53 |
RealBadAngel |
but the code for it is IGUI element copy |
06:53 |
hmmmm |
instead of having drawStatbar() and drawHotbar() and all these little doodads we treat them like pseudo-GUI widgets |
06:53 |
est31 |
also, I think when drawing hud elements in gui, they should be somehow inside the gui's coordinate system |
06:53 |
hmmmm |
each have a Draw |
06:53 |
RealBadAngel |
all the matrices operations come from it |
06:54 |
hmmmm |
the Mapper might be better renamed to Minimap or something |
06:54 |
hmmmm |
and that'll inherit IHudElement |
06:54 |
hmmmm |
or something..? |
06:54 |
hmmmm |
class Minimap : public HudElement { |
06:54 |
hmmmm |
void draw(); |
06:55 |
hmmmm |
and I can't think of any other at the moment |
06:55 |
hmmmm |
i'm sure i'll get a better idea once i start coding it |
06:55 |
RealBadAngel |
hmmmm, my idea was to fully implement it as IGUI element, not our own |
06:55 |
Hunterz |
generated image from the minimap can be used for map item like in the minecraft... |
06:55 |
hmmmm |
RealBadAngel, please don't |
06:55 |
hmmmm |
this is more flexible if it could be rendered anywhere |
06:56 |
RealBadAngel |
the very same way we will be able to display actual cubes in formspecs |
06:56 |
RealBadAngel |
instead of generated textures |
06:56 |
hmmmm |
maybe you could have an IGUIElement for Minimap that does nothing but adds irrlicht GUI nonsense on top of a drawn minimap |
06:56 |
est31 |
yea good idea |
06:56 |
hmmmm |
yes |
06:56 |
hmmmm |
okay |
06:56 |
hmmmm |
so every HUD element will have a corresponding IGUIElement |
06:57 |
est31 |
well it doesnt make much sense to render the inventory bar in gui |
06:57 |
RealBadAngel |
this way we can say goodbye to lotsa old code incuding item previews |
06:57 |
hmmmm |
est31: it doesn't make sense until you need it |
06:57 |
RealBadAngel |
and have for example rotating cubes in selected inventory slot |
06:58 |
hmmmm |
est31: okay, so maybe nevermind on that requirement. but using this infrastructure you should still be able to render an inventory bar on a gui |
06:58 |
est31 |
hmmmm, I just say if there is additional effort from adding things like the inventory bar, we should only add a set for which we can construct use cases |
06:58 |
hmmmm |
just that it won't necessarily need to be a GUI element |
06:58 |
hmmmm |
but the point is |
06:58 |
RealBadAngel |
indeed, just the way stays here |
06:58 |
hmmmm |
no waid |
06:58 |
hmmmm |
wait* |
06:59 |
hmmmm |
the hotbar is a GUI element |
06:59 |
hmmmm |
it's just that it doesn't use the official irrlicht GUI element interface |
06:59 |
est31 |
its graphical and its a user interface |
06:59 |
hmmmm |
I envision a GUI element where you have multiple boxes with images and some text on it that you can scroll through using the mouse wheel when it has focus (in the form of a gui element) |
07:00 |
hmmmm |
so the hotbar is basically a radio button set |
07:00 |
hmmmm |
except irrlicht doesn't have radio buttons |
07:00 |
hmmmm |
here's our chase |
07:00 |
hmmmm |
chance* |
07:00 |
est31 |
well its a button bare |
07:00 |
est31 |
bar* |
07:01 |
RealBadAngel |
just replace images with children gui elements with the same code as minimap have, showing the cubes |
07:01 |
hmmmm |
we might be able to generalize the hotbar to a radio button set |
07:01 |
est31 |
if we give the flexibility to game creators, we should also give them tools to ensure portability |
07:01 |
est31 |
e.g. what about android |
07:01 |
hmmmm |
wait |
07:01 |
hmmmm |
we're thinking too hard about all this |
07:01 |
hmmmm |
let's code first and see what we come up with then continue this conversation |
07:01 |
est31 |
take the voxus hotbar as example |
07:02 |
hmmmm |
talk is cheap, code isn't |
07:02 |
est31 |
its at the side |
07:02 |
RealBadAngel |
i will code now something to show you something |
07:02 |
RealBadAngel |
oops, too much of something ;) |
07:02 |
est31 |
but on android we already have buttons on the side, no room for hotbar |
07:03 |
hmmmm |
the flexibility of the second-gen hotbar means that you'd be able to select your own geometry |
07:03 |
VanessaE |
est31: so move the buttons? |
07:03 |
hmmmm |
when a hotbar is generalized into a set of buttons where only one can be selected at a time |
07:03 |
est31 |
thats something I can agree with |
07:03 |
VanessaE |
a hotbar down one side, control buttons + Dpad on the other side, seems reasonable? |
07:03 |
hmmmm |
if it's implemented right then in theory you should be able to make a hotbar in the shape of a freaking triangle on the screen |
07:03 |
est31 |
or circle like voxelands |
07:03 |
hmmmm |
or maybe in bubbles on random coordinates |
07:04 |
hmmmm |
that's my end-goal here |
07:04 |
hmmmm |
i want to be able to do that |
07:04 |
hmmmm |
i want to be able to transform the hotbar into something where I have random images all on the screen and I can scroll through them using the mouse wheel |
07:05 |
hmmmm |
then I want to be able to use the same hotbar as a GUI element where I can click on them in addition to scrolling with the mouse wheel |
07:05 |
est31 |
or click on them if you are on android |
07:05 |
hmmmm |
wow |
07:05 |
hmmmm |
this all sounds fantastic |
07:05 |
hmmmm |
I need to get real-life work done though :9 |
07:06 |
hmmmm |
the past couple of days I've been wrestling with gmock |
07:06 |
hmmmm |
what a shitty overcomplicated framework |
07:07 |
hmmmm |
don't ever use it for your own projects |
07:09 |
RealBadAngel |
btw, whos the mtgame team? paramat, pilzadam and ? |
07:10 |
est31 |
its on the wiki somewhere |
07:11 |
est31 |
number is 5 |
07:13 |
kahrl |
http://dev.minetest.net/minetest_game_Development |
07:17 |
RealBadAngel |
theres no paramat in here? |
07:38 |
|
julienrat joined #minetest-dev |
07:39 |
|
julienrat left #minetest-dev |
07:42 |
|
nrzkt joined #minetest-dev |
07:42 |
|
Krock joined #minetest-dev |
07:43 |
|
kaeza joined #minetest-dev |
07:51 |
|
diemartin joined #minetest-dev |
08:04 |
|
Yepoleb joined #minetest-dev |
08:39 |
|
diemartin joined #minetest-dev |
08:46 |
est31 |
hrmm |
08:46 |
est31 |
kahrl, I wonder why a pointer is 0, can you help? |
08:46 |
est31 |
ah it isnt null |
08:47 |
est31 |
still there is something fishy going on |
08:47 |
est31 |
a bug I dont know the cause of |
08:47 |
est31 |
when I store some data, everything is fine |
08:47 |
est31 |
when I retrieve it... bam |
08:48 |
|
Krock joined #minetest-dev |
08:49 |
est31 |
ah I have a suspicion |
08:49 |
est31 |
yea |
08:49 |
est31 |
has been my suspicion |
08:51 |
|
Amaz joined #minetest-dev |
09:15 |
|
jin_xi joined #minetest-dev |
10:00 |
|
H-H-H joined #minetest-dev |
10:52 |
Taoki |
RealBadAngel: :) |
10:53 |
Taoki |
(subtle and probably dumb way of asking about that VBO pull request) |
10:53 |
RealBadAngel |
lol |
10:53 |
RealBadAngel |
not yet :) |
10:53 |
Taoki |
ok :) |
10:53 |
RealBadAngel |
still polishing shader issues |
10:54 |
Taoki |
Aah, so it still has a few bugs? I thought you just didn't find the time to create the pull request itself, which is why I found it a bit odd. |
10:55 |
RealBadAngel |
https://imgrush.com/i6v_pYBosH0s.png |
10:55 |
Taoki |
Wow, cool :o |
10:55 |
RealBadAngel |
i had to find a solution for nontileable textures |
10:57 |
Hunterz |
grass block in the hand will be same as ground? |
10:58 |
Taoki |
Hunterz: Shaders don't seem to work for wielded items yet, sadly. Nor players and Lua entities. |
10:58 |
Taoki |
I wonder when that can be fixed |
10:58 |
Hunterz |
ah |
11:01 |
RealBadAngel |
i have code rdy for that but it have flaws |
11:01 |
RealBadAngel |
sooner or later i will fix that |
11:02 |
|
est31 joined #minetest-dev |
11:13 |
|
proller joined #minetest-dev |
11:18 |
|
sloantothebone_ joined #minetest-dev |
11:28 |
|
julienrat joined #minetest-dev |
11:44 |
|
proller joined #minetest-dev |
11:45 |
RealBadAngel |
est31, #2897 its wip |
11:45 |
ShadowBot |
https://github.com/minetest/minetest/issues/2897 -- Fix relief mapping issues by RealBadAngel |
11:46 |
RealBadAngel |
can you check it? |
11:52 |
|
blaaaaargh joined #minetest-dev |
11:55 |
est31 |
looks good |
11:56 |
est31 |
cheapie, you around? |
12:10 |
est31 |
writing glue code feels so glue-y |
12:56 |
|
SopaXT joined #minetest-dev |
13:01 |
|
TeTpaAka joined #minetest-dev |
13:16 |
|
julienrat joined #minetest-dev |
13:40 |
|
Krock joined #minetest-dev |
13:41 |
|
rubenwardy joined #minetest-dev |
13:45 |
|
julienrat joined #minetest-dev |
14:04 |
|
kilbith joined #minetest-dev |
14:53 |
|
RealBadAngel joined #minetest-dev |
15:00 |
|
kilbith joined #minetest-dev |
15:32 |
|
proller joined #minetest-dev |
15:38 |
|
zat joined #minetest-dev |
15:57 |
|
hmmmm joined #minetest-dev |
16:06 |
est31 |
hmmmm, not polished yet and vast parts untested: #2898 |
16:06 |
ShadowBot |
https://github.com/minetest/minetest/issues/2898 -- Add AreaStore data structure by est31 |
16:06 |
est31 |
comments about concept would be great though |
16:06 |
est31 |
I'll read the logs |
16:09 |
hmmmm |
wow it looks really cool |
16:09 |
hmmmm |
I'll have to take a better look sometime |
16:09 |
hmmmm |
this is great |
16:17 |
|
Hunterz joined #minetest-dev |
16:37 |
|
CraigyDavi joined #minetest-dev |
16:37 |
|
CraigyDavi joined #minetest-dev |
16:46 |
|
RealBadAngel joined #minetest-dev |
16:52 |
|
jin_xi joined #minetest-dev |
17:13 |
|
kilbith joined #minetest-dev |
17:24 |
|
paramat joined #minetest-dev |
17:25 |
paramat |
yes i was wrong to state the memory leak was due to mapgen, i wasn't sure |
17:27 |
|
OldCoder joined #minetest-dev |
17:28 |
|
MinetestForFun joined #minetest-dev |
17:28 |
paramat |
hi sfan5 can i push this fix for yesterday's mesh stair commit? game#561 RealBadAngel please can you check it is correct? |
17:28 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/561 -- Improved stairs model by kilbith2 |
17:36 |
|
est31 joined #minetest-dev |
17:41 |
|
est31 joined #minetest-dev |
17:47 |
|
Icedream joined #minetest-dev |
17:57 |
|
H-H-H joined #minetest-dev |
18:28 |
|
paramat left #minetest-dev |
18:50 |
VanessaE |
RealBadAngel: the screenshot looks good. |
19:07 |
|
hmmmm joined #minetest-dev |
19:17 |
|
selat joined #minetest-dev |
19:56 |
|
paramat joined #minetest-dev |
19:57 |
paramat |
okay thanks sfan will push game 561 later |
20:01 |
|
est31 joined #minetest-dev |
20:06 |
paramat |
perhaps there is not a memory leak at all, but we are just seeing expected world and mesh data accumulating in the client? |
20:09 |
est31 |
if you leave to main menu, the memory doesnt decrease |
20:09 |
est31 |
that makes it a leak |
20:10 |
VanessaE |
that's if you CAN leave to the main menu :P |
20:10 |
paramat |
yes that behaviour is a bug |
20:11 |
VanessaE |
(for me, the game hard-crashes pretty frequently when exiting to main menu) |
20:11 |
est31 |
thats pretty bad |
20:11 |
est31 |
even horrible |
20:11 |
est31 |
I have it on a certain server |
20:21 |
|
crazyR joined #minetest-dev |
20:31 |
|
Warr1024 joined #minetest-dev |
20:34 |
paramat |
now pushing game#561 |
20:34 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/561 -- Improved stairs model by kilbith2 |
20:38 |
paramat |
complete |
20:47 |
paramat |
so mesh stairs are now fixed. RealBadAngel the only problem with defining stairs as meshes is they appear dark as a wielditem, is there a fix for this? |
20:48 |
VanessaE |
paramat: that seems to happen with any kind of mesh |
20:48 |
VanessaE |
including extruded meshes and the ones you see from the wield3d mod |
20:49 |
paramat |
okay, i know our wield is a bit buggy at the moment |
21:06 |
est31 |
hmmmm, do you know something about templates https://github.com/est31/minetest/commit/6b85607215d70faf7ce8e58c0b51c83c66d6820c#diff-e5ab4d3af8e375631903ed35b243c3b4R312 |
21:07 |
est31 |
I want to implement a simple LRU cache |
21:07 |
est31 |
with templates |
21:07 |
est31 |
so I've chosen the simplest approach: direct definition in header |
21:07 |
est31 |
but doesnt build |
21:07 |
hmmmm |
don't use templates |
21:08 |
est31 |
it sais a ton of errors |
21:08 |
hmmmm |
generally they suck |
21:08 |
est31 |
they do indeed |
21:08 |
hmmmm |
yea |
21:08 |
hmmmm |
and the errors are impossible to figure out |
21:09 |
hmmmm |
I think the actual problem you have is that you can't have nested templated types |
21:09 |
est31 |
templates do have their use though |
21:09 |
est31 |
? |
21:09 |
hmmmm |
yeah, for language experts whose goal is to make a library and ugly code |
21:10 |
est31 |
util/container.h is very library-ish |
21:10 |
hmmmm |
you need to define std::list<K> and so on along with the template definition |
21:10 |
est31 |
like? |
21:12 |
VanessaE |
paramat: what say you about adding other basic shapes to minetest_game's models selection? like slopes, slabs, other shapes as in my slope_test mod? |
21:12 |
VanessaE |
i.e. things other mods can use instead of carrying their own copies |
21:12 |
hmmmm |
template <class K, class KeyList = std::list<K>, class KeyValuePair = std::pair<std::list<K>::iterator, V> > > |
21:13 |
hmmmm |
? |
21:13 |
hmmmm |
I think |
21:13 |
paramat |
only very basic diagonals, because Minetest already has diagonal plantlike drawtype as part of it's character |
21:13 |
hmmmm |
every time I thought to myself, "gee, templates would really work nicely here" it ended up not working nicely and I trashed it |
21:14 |
paramat |
so perhaps any shape that uses surfaces created from 45 degree diagonals between node vertices |
21:15 |
paramat |
a bit like when LEGO added 45 degree roofing slopes =) |
21:15 |
paramat |
(well not 45 degree actually) |
21:16 |
VanessaE |
paramat: yeah, I was thinking the same |
21:17 |
paramat |
basic slope, inner corner, outer corner |
21:17 |
VanessaE |
like the three basic roof shapes in homedecor (they're all 45-degree angles) |
21:17 |
VanessaE |
exactly that |
21:18 |
VanessaE |
LEGO has 2:1 and 3:1 slopes yeah, but I think they also have 1:1 |
21:19 |
VanessaE |
paramat: there are a few shapes in Moreblocks that meet your criteria also btw |
21:19 |
VanessaE |
(corner pyramids and the like) |
21:20 |
VanessaE |
but that's overkill |
21:20 |
VanessaE |
wait...don't I normally suggest overkill anyway? :D |
21:20 |
paramat |
yes heh |
21:21 |
VanessaE |
anyways, there are lots of primitives in slope_test, technic CNC machine, and moreblocks that you'll want to pull in, I'm sure |
21:27 |
paramat |
weird, now i'm getting big drops in memory use when exiting a world, as if most (perhaps not all) world data is now being correctly deleted |
21:27 |
paramat |
in very latest dev |
21:30 |
paramat |
well i'm bemused and may have been doing something silly. i'll put this aside now i have other stuff to focus on |
21:43 |
|
sloantothebone joined #minetest-dev |
21:44 |
|
sloantothebone joined #minetest-dev |
21:44 |
paramat |
uh-oh |
21:45 |
VanessaE |
? |
21:47 |
|
selat joined #minetest-dev |
21:54 |
|
RealBadAngel joined #minetest-dev |
22:07 |
est31 |
yaay it compiles |
22:08 |
est31 |
or does it |
22:08 |
VanessaE |
est31: "It compiles, ship it!" :) |
22:08 |
est31 |
heh |
22:09 |
est31 |
I wonder how the cache will affect the benchmarks |
22:09 |
est31 |
wow it *actually* compiles |
22:13 |
|
CraigyDavi joined #minetest-dev |
22:15 |
|
troller joined #minetest-dev |
22:26 |
est31 |
so why does the cache crash? |
22:28 |
est31 |
ah |
22:28 |
est31 |
order :p |
22:43 |
paramat |
sfan5 and all, game#562 |
22:43 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/562 -- Default: New ice texture by Gambit by paramat |
22:55 |
Taoki |
RealBadAngel: Are you still around? |
23:12 |
VanessaE |
Taoki: so, when you gonna make the model's head able to yaw/pitch separately from the body when looking around? :) |
23:12 |
Taoki |
VanessaE: When Irrlicht stops breaking models that enable manual IK control. |
23:12 |
Taoki |
That's the right way to do it |
23:12 |
VanessaE |
heh |
23:12 |
est31 |
IK? |
23:18 |
Taoki |
est31: Putting the armature in a mod which allows controlling bone positions manually |
23:18 |
Taoki |
That breaks animations |
23:49 |
est31 |
irrlicht has many bugs |
23:50 |
est31 |
unfortunately |
23:50 |
est31 |
age doesnt guarantee bug-freeness |