Time |
Nick |
Message |
00:14 |
|
Terusthebird joined #minetest-dev |
00:49 |
|
sapier1 joined #minetest-dev |
00:50 |
|
sapier1 left #minetest-dev |
03:42 |
|
Terusthebird joined #minetest-dev |
03:46 |
|
paramat joined #minetest-dev |
03:47 |
paramat |
will merge https://github.com/minetest/minetest/pull/3456 'Mgfractal: Add 3D and 4D fractals' in 20+ mins |
04:15 |
paramat |
now merging |
04:23 |
paramat |
merged |
04:28 |
|
paramat left #minetest-dev |
04:34 |
|
eeew joined #minetest-dev |
05:08 |
sofar |
can someone explain to me why the following patch works for X++ but not for Z++: |
05:08 |
sofar |
https://gist.github.com/sofar/c04e2932967cb9edb768 |
05:08 |
sofar |
connects fencelike drawtype to normal drawtype |
05:09 |
sofar |
it never paints the bars to the Z++ direction normal node |
05:09 |
sofar |
but works fine for the X++ direction |
05:11 |
sofar |
so weird |
05:12 |
sofar |
wheps, that totally works fine |
05:13 |
sofar |
must have forgotten to build? |
05:19 |
sofar |
ahhh what am I getting myself into |
05:20 |
VanessaE |
you dived into the engine source. you're screwed now ;) |
05:21 |
sofar |
if I could compose a xrailuv array by head this patch would be done |
05:21 |
sofar |
it's slightly more complex, don't want to have the bars extend into the normal node... |
05:21 |
sofar |
sigh |
05:21 |
VanessaE |
it might be a good idea to introduce a node property that allows a fence to connect to it when it otherwise isn't a "normal" node. |
05:21 |
sofar |
I just don't know where to start with this game |
05:21 |
VanessaE |
(e.g. something made of nodeboxes or a mesh that's still mostly cubical) |
05:23 |
sofar |
which incedentally are also needed for e.g. itemframes and torches |
05:23 |
sofar |
can_be_mounted_to |
05:23 |
VanessaE |
mmhmm |
05:24 |
sofar |
ugh this sea mod is terrible, hello kitty island adventure meets nyan cat colors in my unified_inventory |
05:24 |
sofar |
can't find a damn node |
05:25 |
VanessaE |
haha |
05:26 |
sofar |
so I started out looking what the problem with underwater plants is |
05:26 |
VanessaE |
could be worse: imagine coloredwood mod with moreblocks installed at the same time. together they generate something like 3500 nodes (!). |
05:26 |
sofar |
not sure how I got derailed into looking at connecting fences to normal nodes |
05:26 |
VanessaE |
(thankfully coloredwood isn't in the creative inv :) ) |
05:26 |
VanessaE |
wellm |
05:26 |
VanessaE |
well, |
05:27 |
VanessaE |
it's like a youtube or wikipedia clickfest |
05:27 |
VanessaE |
one moment you're looking at the article on screwdrivers, and 30 mins later you've somehow ended up on an article about cuttlefish. |
05:27 |
sofar |
my stash of modifications to default is growing as well :/ |
05:27 |
sofar |
lockable trapdoors |
05:27 |
VanessaE |
submit them as PRs. maybe they'll be useful to others. |
05:27 |
sofar |
better looking grass |
05:28 |
sofar |
keys to lockable objects |
05:28 |
sofar |
(can really only put that in minetest_game, as a mod won't work) |
05:28 |
sofar |
a recipe to copy books |
05:29 |
sofar |
colored carpet |
05:29 |
sofar |
tossing out spores for mushrooms |
05:29 |
sofar |
(seriously?) |
05:29 |
DFeniks |
you know what i always wanted? horizontal fence , so i can build something like axis for wheels |
05:30 |
sofar |
of course I've patched in 3d torches with nice particles |
05:30 |
DFeniks |
maybe it even exists? already |
05:30 |
VanessaE |
spores for mushrooms make at least some sense, since they were also in some other mushroom mods. |
05:30 |
VanessaE |
DFeniks: a nodebox the size of a fence, turned sideways, would accomplish that |
05:30 |
sofar |
not a single mushroom grower uses spores |
05:30 |
VanessaE |
easily done with amod. |
05:31 |
DFeniks |
yeah |
05:31 |
sofar |
they spread, you pick all but a few, they spread, you pick all but a few, they spread |
05:31 |
sofar |
fences for the 3 other wood types |
05:31 |
sofar |
(that's like 3 lines) |
05:31 |
VanessaE |
one thing I'd like to see are *diagonal* fences (between any two X/Z-adjacent coords that is) |
05:32 |
VanessaE |
(darkrose's fork has such) |
05:32 |
sofar |
that's possible both in core and in a mod |
05:32 |
DFeniks |
but im thinking on servers where i play , i would need to make mod and make them install it |
05:32 |
sofar |
I'd rather see walls (Oh hey I made my own wall mod) |
05:32 |
sofar |
NDT_WALLLIKE |
05:33 |
VanessaE |
DFeniks: alternately, if a server has streets mod, then there are probably the standard utility poles that it comes with, which can be turned sideways also. |
05:34 |
VanessaE |
yeah, diagonal fences can be done with a mod, but that means adding several nodes. |
05:34 |
sofar |
needs to be a meshnode, yes |
05:34 |
sofar |
very awkward |
05:35 |
sofar |
I also made a library mod so I can store my books in a lua file |
05:35 |
sofar |
e.g. /library add |
05:35 |
VanessaE |
nice. |
05:35 |
sofar |
adds book in hands |
05:35 |
sofar |
you can /library give #number |
05:35 |
sofar |
or /library remove #number |
05:35 |
sofar |
list |
05:35 |
DFeniks |
actually , did i ever seen mesh node thats can be rotated? althought probably slopes |
05:36 |
sofar |
mesh nodes can be rotated |
05:36 |
sofar |
easily |
05:36 |
VanessaE |
yeah, they work like any other facedir node |
05:36 |
* DFeniks |
mostly builds with slabs stairs and fences |
05:37 |
sofar |
k, maybe I'll try and push some stuff from my queue to minetest_game |
05:49 |
sofar |
VanessaE: https://github.com/minetest/minetest_game/pull/742 |
05:49 |
sofar |
there be first one of tonight |
05:50 |
VanessaE |
nice idea |
05:51 |
sofar |
plus, and this took me a while, that shows how you can craft multiple items at the same time |
05:52 |
VanessaE |
I see it. |
05:52 |
VanessaE |
(I think) |
05:52 |
sofar |
frankly others have done that before, but I could never find an example in minetest_game |
05:52 |
sofar |
this copy method works really well to copy one book a lot of times |
05:53 |
sofar |
you can put 99 empty books and 1 written one in the crafting table and just click, click, click |
06:15 |
|
leat joined #minetest-dev |
06:25 |
|
RealBadAngel joined #minetest-dev |
06:31 |
RealBadAngel |
sofar, but that patch wont be enough, NDT_NORMAL wont connect to fences, you need to add X- and Z- too |
06:32 |
sofar |
RealBadAngel: yes, but also we don't want to make bars that poke a hole into the normal node |
06:32 |
RealBadAngel |
you need to add half of the bar |
06:32 |
sofar |
so it needs 4 new if (side is normal) { draw bar } sections |
06:32 |
sofar |
exactly |
06:32 |
sofar |
other than that it should be a nice visual fix for fences |
06:36 |
VanessaE |
sofar: see also game #665 |
06:37 |
VanessaE |
(my comment at the end) |
06:38 |
sofar |
note I replied on that one myself too |
06:38 |
VanessaE |
I saw. |
06:39 |
sofar |
I don't like his craft recipes |
06:39 |
sofar |
mine just adds acacia sticks and whatnot :^) |
06:39 |
sofar |
(who cares what sticks they were |
06:39 |
sofar |
plus I like the clutter it creates) |
06:39 |
* VanessaE |
looks |
06:39 |
sofar |
I don't have it in git yet |
06:40 |
VanessaE |
hm |
06:40 |
VanessaE |
actually, |
06:40 |
VanessaE |
that's not a bad recipe idea |
06:40 |
sofar |
I made my API a bit cleaner, I think |
06:40 |
sofar |
register_fence(name, {def}) |
06:40 |
sofar |
plus I add the sticks needed (and textures for those) |
06:48 |
|
Hunterz joined #minetest-dev |
07:22 |
|
GhostDoge joined #minetest-dev |
07:27 |
|
Obani joined #minetest-dev |
07:45 |
sofar |
https://github.com/minetest/minetest_game/pull/744 may be the least contentious addition for now... |
07:48 |
sofar |
VanessaE: one particularly core issue I'd like to see solved is underwater plantlike nodes. now that I wish I had an idea on how to fix (only have seen every underwater mod struggle with hacks to solve drawing issues) |
07:49 |
everamzah |
mm, water torches |
07:49 |
VanessaE |
*nod* something like that in-engine could be useful but it would be tricky |
07:49 |
VanessaE |
(particularly where water surface shaders are concerned) |
07:50 |
sofar |
I'm wondering if builtin can handle the on_place in water |
07:50 |
sofar |
but likely it should not |
07:50 |
sofar |
also, underwater plants gonna dry out if the water disappears? |
07:50 |
sofar |
not easy to solve |
07:52 |
VanessaE |
that's easy. turn it into dry shrub :) |
07:52 |
sofar |
yes but it'll be in a bubble of fluid |
07:54 |
VanessaE |
not if you replace it with an ordinary dry shrub node |
07:55 |
sofar |
true, if the node being "under water" is being handled in content_mapblock.cpp properly |
07:55 |
sofar |
then an ABM could do just that |
07:56 |
VanessaE |
well that's something you'd want the engine to switch out the instant the water "runs out" e.g. during the water spread code |
08:01 |
sofar |
yeah, if there's a flag somewhere that tells the engine to draw this like a normal underwater node, but it's not actually "in a fluid" thent that solves that. |
08:05 |
sofar |
i think I've ran out of easily mergable minetest_game patches |
08:05 |
sofar |
the rest gets more complicated, like keys |
08:33 |
|
Darcidride joined #minetest-dev |
08:34 |
|
Darcidride joined #minetest-dev |
08:35 |
|
dzho joined #minetest-dev |
08:38 |
sofar |
http://i.imgur.com/7DYxc4a.png |
08:38 |
sofar |
almost done |
08:38 |
sofar |
just need to fix the UV's |
08:38 |
sofar |
but, it's late |
09:05 |
|
jin_xi joined #minetest-dev |
09:16 |
|
nrzkt joined #minetest-dev |
09:56 |
|
DFeniks_ joined #minetest-dev |
10:13 |
|
VargaD joined #minetest-dev |
10:21 |
|
VargaD joined #minetest-dev |
11:59 |
|
hmmmm joined #minetest-dev |
12:28 |
|
behalebabo joined #minetest-dev |
12:32 |
|
rubenwardy joined #minetest-dev |
12:36 |
|
Fixer joined #minetest-dev |
12:59 |
|
zat joined #minetest-dev |
13:20 |
|
gravgun joined #minetest-dev |
13:50 |
|
est31 joined #minetest-dev |
13:51 |
est31 |
~tell dfelinto there is an 3d_mode setting for the client, perhaps you'll want that. |
13:51 |
ShadowBot |
est31: O.K. |
14:02 |
|
Calinou joined #minetest-dev |
14:45 |
|
Player2 joined #minetest-dev |
15:09 |
|
Obani joined #minetest-dev |
15:32 |
|
kaeza joined #minetest-dev |
15:33 |
|
CraigyDavi joined #minetest-dev |
15:53 |
est31 |
there seems really no other way to fix the keybindings problem on russian keyboards than to patch irrlicht |
15:53 |
est31 |
this would mean that we ship our own copy of the engine |
15:53 |
est31 |
hehe |
15:53 |
est31 |
or to propose it for 1.9, and then download irrlicht automatically at the build step |
15:54 |
est31 |
(as the distros wont package a version of irrlicht that isnt officially "released") |
15:55 |
nrzkt |
irrlicht 1.9 has some crash on some memset |
15:55 |
nrzkt |
i'm talking with devs about this |
15:55 |
est31 |
in forums? |
15:55 |
est31 |
or irc |
15:55 |
nrzkt |
but yes, it really fix the keyboard issues, i can confirm that my special minetest declinaison fix problems |
15:56 |
nrzkt |
i'm talking on IRC, but i need to fix some problems to have exactly the piece of code working instead of the whole project |
15:57 |
est31 |
1.9 fixes them you say? |
15:57 |
est31 |
in what way |
15:58 |
nrzkt |
i don't know in what way but it fixes all problems i have with my french keyboard |
15:58 |
est31 |
there is this obvious mistake with the event keycode not sending an wchar but some utf8 encoded thingy |
15:58 |
nrzkt |
numpad, + 7 not working + accents etc |
15:58 |
nrzkt |
and i can connect to my server |
15:58 |
nrzkt |
but client crash when loading water block texture |
15:59 |
est31 |
and the other mistake is that the other keycode doesnt work for international keyboads |
16:00 |
est31 |
they have to apply the keyboard mapping backwards or so |
16:00 |
est31 |
right now they take the least significant byte I think of the wchar |
16:00 |
est31 |
totally wrong ofc :) |
16:00 |
est31 |
so nrzkt you say you can bind 7 for movement? |
16:00 |
est31 |
e.g. set 7 to move forward |
16:04 |
nrzkt |
7 has no effect with current irrlicht/MT |
16:04 |
nrzkt |
with irrlicht 1.9 7 works like it does |
16:04 |
nrzkt |
and all numpad too |
16:19 |
Calinou |
we should bundle Irrlicht perhaps |
16:19 |
Calinou |
to fix iti ASAP |
16:19 |
Calinou |
-i |
16:19 |
Calinou |
after the bugs are fixed |
16:31 |
|
gravgun joined #minetest-dev |
16:33 |
Darcidride |
(someoen told me one time, (i did'nt remember who) he compilates minetest with irrlicht 1.9, it works well for him and he doesn't have anymore the binding problem) |
16:33 |
est31 |
ha I killed Calinou |
16:35 |
est31 |
oh offtopic |
16:49 |
|
alket joined #minetest-dev |
16:55 |
|
Hunterz joined #minetest-dev |
17:04 |
|
nrzkt joined #minetest-dev |
17:07 |
|
jin_xi joined #minetest-dev |
17:11 |
|
zupoman joined #minetest-dev |
17:11 |
|
zupoman joined #minetest-dev |
17:17 |
|
troller joined #minetest-dev |
17:28 |
|
Krock joined #minetest-dev |
17:43 |
|
Robert_Zenz joined #minetest-dev |
17:45 |
|
BlockMen joined #minetest-dev |
17:59 |
|
BlockMen joined #minetest-dev |
18:08 |
Fixer |
Calinou, or unbundle ._. |
18:19 |
sofar |
VanessaE: did you see http://i.imgur.com/7DYxc4a.png ? |
18:19 |
|
Calinou joined #minetest-dev |
18:20 |
Krock |
Bugged visual _< |
18:20 |
Krock |
*:< |
18:20 |
sofar |
Krock: as said, I need to fix the uvmaps, then it should be fine |
18:21 |
Krock |
oh good to hear ^^ |
18:21 |
sofar |
I've never done uvmaps other than in blender, so I didn't have the time before bedtime :^) |
18:21 |
gravgun |
Meanwhile, in a world with borked Z-ordering... |
18:21 |
Krock |
> broked |
18:21 |
Krock |
Lol I even wrote the comment wrong |
18:21 |
gravgun |
Real english is "broken" |
18:21 |
Krock |
yeah. |
18:22 |
sofar |
Krock: should be trivial to fix, but I didn't want to do a half-ass job at bad uvmaps, so I might take some time to fix that |
18:22 |
gravgun |
http://www.urbandictionary.com/define.php?term=borked |
18:22 |
Fixer |
i wonder if "buggen" is possible |
18:22 |
sofar |
plus I'm going to be travelling for a week or so |
18:22 |
|
BlockMen joined #minetest-dev |
18:22 |
Fixer |
sofar, but will they accept that PR? |
18:22 |
gravgun |
We don't know so far |
18:22 |
gravgun |
*ba dum tsss* |
18:23 |
sofar |
the alternative is that everyone has to make nodeboxes in mods |
18:23 |
* gravgun |
shows himself the door |
18:23 |
sofar |
which is slow, and proliferates bad implementations |
18:23 |
sofar |
we have NDT_FENCELIKE |
18:23 |
Fixer |
last what i've seen is dev defence against that feature |
18:23 |
sofar |
if it's broken now, we should fix it |
18:23 |
Krock |
gravgun, I wonder where all those slangs come from o.o |
18:23 |
sofar |
or just remove it |
18:24 |
gravgun |
Krock: the interwebs are very, very vast. Much like slang emerges in towns, so does it on the net. |
18:25 |
BlockMen |
sofar, better provide a common way to define fences |
18:26 |
BlockMen |
else we need a second type for stonewalls |
18:26 |
BlockMen |
and a third for xyz |
18:26 |
sofar |
BlockMen: agreed, very much so |
18:26 |
sofar |
BlockMen: of course that will have everyone up in arms even more so |
18:27 |
sofar |
stonewalls are nice, I made some myself. all wood materials -> fence, all stone materials -> stone |
18:27 |
sofar |
s/wall |
18:27 |
sofar |
makes for nice villages in theme of local resources |
18:29 |
BlockMen |
sofar, i did that with a new nodebox type, but dont have the code anymore: https://www.youtube.com/watch?v=ouTtZ1UFDCA :\ |
18:29 |
BlockMen |
so if you want to reimplement, feel invited ;) |
18:29 |
VanessaE |
sofar: I did now. :) that's very much like what I was thinking of. |
18:30 |
VanessaE |
(the screenshot) |
18:30 |
sofar |
BlockMen: lulz, fence gate |
18:31 |
BlockMen |
sofar : huh? |
18:32 |
sofar |
BlockMen: yeah, absolutely, a better NDT_FENCELIKE would be a major improvement |
18:33 |
|
troller joined #minetest-dev |
18:33 |
VanessaE |
sofar: acacia, jungle, pine, default wood, .... what material is that fence on the far right? |
18:33 |
BlockMen |
no, not NDT_FENCELIKE. we already have a drawtype for costum shapes. just add a new type (beside leveled and fixed) |
18:34 |
sofar |
VanessaE: modified nether mod - added nether fence/brick/stairs |
18:34 |
VanessaE |
ah. |
18:34 |
sofar |
BlockMen: would solve compatibility going back/forward, sure |
18:35 |
sofar |
how about NDT_CONNECTING - something that could be used for e.g. mese circuits, xpanes, etc? |
18:35 |
sofar |
VanessaE: https://github.com/PilzAdam/nether/pulls |
18:35 |
BlockMen |
sounds fine for me^ |
18:36 |
BlockMen |
* not the link, your suggestion |
18:36 |
sofar |
sure |
18:36 |
sofar |
is pilzadam on a sabbatical or something, btw? |
18:36 |
VanessaE |
he's here randomly |
18:39 |
sofar |
28 days open pull request on his github :/ |
18:40 |
sofar |
one PR open for 6+ months now |
18:40 |
sofar |
!seen pilzadam |
18:40 |
ShadowBot |
sofar: I saw pilzadam in #minetest-dev 5 weeks, 1 day, 22 hours, 14 minutes, and 4 seconds ago saying "my commit count doesn't complain either ;-)" |
18:43 |
|
H-H-H joined #minetest-dev |
18:52 |
|
blaze joined #minetest-dev |
19:02 |
|
est31 joined #minetest-dev |
19:02 |
est31 |
sofar, technic cables would profit from such a mod/change too |
19:02 |
est31 |
just wondering why it shouldn't be nodeboxes |
19:03 |
sofar |
could be just planes as well |
19:03 |
sofar |
client engine can make them far easier without node proliferation |
19:03 |
sofar |
mods have to rotate them and make all sorts of variants |
19:03 |
est31 |
you mean these ugly things that can be observed only if looked at from the right angle? |
19:04 |
est31 |
like plants or torches |
19:06 |
sofar |
no |
19:06 |
sofar |
I mean having to define all wall variants as nodes |
19:06 |
sofar |
whereas the core engine can do it with 1 node |
19:06 |
est31 |
yeah |
19:07 |
est31 |
technic cables are about 40 nodes i think |
19:08 |
sofar |
exactly that |
19:08 |
sofar |
would all go away |
19:08 |
est31 |
yeah |
19:32 |
|
younishd joined #minetest-dev |
19:34 |
|
sapier joined #minetest-dev |
20:02 |
|
younishd joined #minetest-dev |
20:24 |
|
Vexyl joined #minetest-dev |
20:53 |
|
RealBadAngel joined #minetest-dev |
20:53 |
|
Darcidride joined #minetest-dev |
21:08 |
RealBadAngel |
is plural from dependecy "dependiences" or "dependencys" > |
21:08 |
RealBadAngel |
? |
21:09 |
VanessaE |
dependencies |
21:09 |
RealBadAngel |
https://github.com/minetest/minetest/pull/3454 |
21:10 |
sapier |
why do I always change titles on request of someone just to realize it's been ok before? |
21:10 |
est31 |
its dependencies yes |
21:10 |
RealBadAngel |
hehe |
21:11 |
BlockMen |
sapier use blank titles :P |
21:11 |
RealBadAngel |
its highlight of the most common way here, comment typos, whities not the code :P |
21:12 |
RealBadAngel |
btw, sapier, howdy pal |
21:12 |
RealBadAngel |
where have u been, what have u been doing all this time? |
21:12 |
sapier |
needed a break |
21:12 |
RealBadAngel |
i took one too |
21:13 |
sapier |
and my other hobbies needed some attention too ... doing 3d-printing takes a lot of time |
21:14 |
BlockMen |
hmmm...probier mal beide zu kombinieren |
21:15 |
est31 |
3d printing minetest scene |
21:15 |
est31 |
minetest 3d printer mod |
21:16 |
est31 |
build it in minetest, it gets then printed by the printer |
21:17 |
BlockMen |
sound like the question of what was first: chicken or egg |
21:19 |
est31 |
egg came first, en had sex with chick, chick laid egg out came chicken |
21:19 |
sofar |
I've been thinking of making a minetest mod that allows me to virtually manipulate openstack nodes |
21:19 |
est31 |
bad joke |
21:19 |
RealBadAngel |
im makin now 3d texture nodes |
21:19 |
RealBadAngel |
kinda related ;) |
21:19 |
sofar |
what is "3d texture" ? |
21:20 |
RealBadAngel |
i tried before but i died with plants waving |
21:20 |
sofar |
oh you explained to me earlier |
21:20 |
BlockMen |
i guess he means extruding, no? |
21:20 |
RealBadAngel |
now i do have each pixel to be 1/16th of the full node size |
21:20 |
sofar |
no, 16x16x16 pixels in a node |
21:21 |
RealBadAngel |
i mean 16 layers of texture |
21:21 |
BlockMen |
ah, ok |
21:21 |
RealBadAngel |
each layer describing 16x16 layer |
21:21 |
RealBadAngel |
this way inside one node we do have 256 mininodes |
21:22 |
RealBadAngel |
most simple form is "extruded" plantlike |
21:22 |
RealBadAngel |
most advanced is l-system plant using it |
21:25 |
RealBadAngel |
and such node is going to be freely used by mods |
21:26 |
RealBadAngel |
i mean ability to change a 3d pixel runtime |
21:27 |
RealBadAngel |
for sake of simplicty i decided to make such pixel be single coloured, not another node in node |
21:28 |
RealBadAngel |
this way we can make everything we already have (but flat) and make it all 3d |
21:38 |
RealBadAngel |
anyway, this is kinda long time, i need about 2-3 weeks more to fininish it |
21:39 |
RealBadAngel |
i do have basics working, but fine tuning is damn hard |
21:39 |
sapier |
hmm using latest android sdk requires to build for at least android-15 (4.0) atm our minimum requirement is android-10 (2.3) |
21:40 |
RealBadAngel |
sapier, dont you think we shall include (and fork) irr once and for all? |
21:40 |
sapier |
no for sure not |
21:40 |
RealBadAngel |
im listening for cons |
21:40 |
RealBadAngel |
(tux did it) |
21:41 |
sapier |
I don't know if it's just for the last days but at least those days it doesn't look like we have tripled our developer count |
21:41 |
sapier |
we just don't have manpower to create and more important maintain patches |
21:41 |
nrzkt |
sapier, 3.5% of current minetest app users are using android 2, but for every install not only current installation 11% are android 2.3 |
21:41 |
RealBadAngel |
fightinin with incompatible (or broken) irr releases doesnt help |
21:42 |
sapier |
I just ported the texture hack for irrlicht and althou this was a 20 line patch it took hours to find out how to match it to recent irrlicht changes |
21:42 |
est31 |
nrzkt, can you upload screenshot of stats somewhere? |
21:42 |
RealBadAngel |
1.8.1 is common and broken |
21:42 |
RealBadAngel |
at this point we can grab workin one and get it |
21:42 |
RealBadAngel |
and forget bout irr related problems... |
21:43 |
nrzkt |
est31: https://lut.im/UiP5O8s59L/Fa9WEFxWsA8Hqxlh.png & https://lut.im/fuzTsQbeN4/RS0jL6K7QoM7OPyO.png |
21:43 |
sapier |
as I don't see a benefit from switching to recent sdk I suggest staying where we are for the moment |
21:43 |
nrzkt |
irrlicht 1.8.3 should be used, not 1.8.1 |
21:43 |
RealBadAngel |
sapier, benefit is to stick to working one |
21:43 |
RealBadAngel |
for starters |
21:43 |
sapier |
irrlicht is used in the version provided by distribution |
21:44 |
RealBadAngel |
and that way is fairly wrong |
21:44 |
sapier |
RealBadAngel: sdk version doesn't have much influence on minetest |
21:44 |
sapier |
you can even use recent ndk (which does have influence) and combine it with old sdk |
21:44 |
RealBadAngel |
we shall pick WORKIN version |
21:44 |
est31 |
yeah |
21:44 |
sapier |
most of minetest is native code that's why I mention it |
21:45 |
est31 |
I'm with sapier switching sdk needs a reason |
21:45 |
RealBadAngel |
not the distromaintainers thought and assumed to be workin |
21:45 |
est31 |
and right now there is none |
21:45 |
sapier |
I never saw any compatibility issues with any version |
21:45 |
RealBadAngel |
because we do now that some of them are NOT workin |
21:45 |
RealBadAngel |
1.8.1 |
21:45 |
sapier |
est31, primary reason would be availability AND gcc version |
21:45 |
nrzkt |
est31, does those images fits to you ? |
21:45 |
RealBadAngel |
next plz |
21:45 |
sapier |
recent gcc usually has less bugs |
21:45 |
sapier |
and does build more optimized code |
21:46 |
est31 |
nrzkt, yeah, they are good. thanks! |
21:47 |
sapier |
RealBadAngel: for distributions irrlicht version ain't our decision and on android it's just a matter of deciding between bad options as all ogles irrlicht versions are work in progress |
21:47 |
nrzkt |
sapier, with clang 3.6 and 3.7 the gap between gcc and clang is a little bit less hig |
21:47 |
est31 |
sapier, didn't know that the gcc version is fixed for target sdk, interesting |
21:47 |
est31 |
RealBadAngel, the non working irrlicht version is 1.8.2 btw |
21:47 |
RealBadAngel |
sapier, when you choose workin one as a platform, you can merge fixes you need |
21:47 |
sapier |
nope est31 gcc version is part of ndk which I mentioned to be changeable without issues |
21:48 |
est31 |
ah I see |
21:48 |
RealBadAngel |
but im not the one to decide |
21:48 |
RealBadAngel |
im just talkin about the way to avoid some problems, common ones |
21:48 |
sapier |
nrzkt: did you manage to compile it using clang? I never tried but there's at least beginning of support in it |
21:48 |
nrzkt |
1.8.3 or 1.8.1 are good versions, 1.9 is not released and has some anoying crashes. 1.8.2 is buggy |
21:48 |
RealBadAngel |
we have tried including irr and it works nicely |
21:48 |
nrzkt |
sapier, i only build with clang on FreeBSD adding more warnings |
21:49 |
nrzkt |
we have a FreeBSD builder on my jenkins using clang 3.6 |
21:49 |
nrzkt |
or 3.4 |
21:49 |
nrzkt |
if you look at your PR there is a FreeBSD/clang builder :) |
21:49 |
sapier |
well I meant android, there's a clang option too |
21:49 |
RealBadAngel |
and theres a reason why tux kart team did it |
21:49 |
nrzkt |
oh |
21:49 |
sapier |
but I don't know if our dependencies build with clang |
21:49 |
nrzkt |
i built only with gcc for the play store |
21:50 |
nrzkt |
with the last sdk/ndk i think because i'm on archlinux |
21:50 |
nrzkt |
i have all sdk from 2.3 to 6.0 because i'm developping some android apps |
21:51 |
sapier |
Still my suggestion is stay at current sdk version and update ndk as far as possible |
21:51 |
est31 |
as long as it works, fine with me |
21:51 |
nrzkt |
i do'nt see a reason to upgrade SDK/NDK for android. We don't use specific things in our code linked with android |
21:52 |
nrzkt |
SDK version is only important for API in the android interface, that we don't care |
21:53 |
sapier |
exactly but ndk does provide better compiler and as competition between gcc and clang recently resulted in improvements in both of them I'd hope it might help a little bit |
21:54 |
nrzkt |
yes |
21:54 |
est31 |
it did? |
21:54 |
est31 |
didnt know |
21:54 |
est31 |
market rules even working in open source world |
21:54 |
est31 |
horray |
21:54 |
nrzkt |
didn't know and the compiler optimisations will not fix our design which is not adapted for android :) |
21:56 |
sapier |
I'm trying gcc-4.9 instead of 4.8 now |
21:56 |
est31 |
sapier, what precisely is that texture hack about? |
21:56 |
sapier |
there are gpu's out claiming to support GL_BGRA textures but don't really do it |
21:57 |
sapier |
e.g. my sony xperia M device does |
21:57 |
sapier |
irrlicht guys download texture to gpu but never ever check for ogl errors |
21:57 |
RealBadAngel |
with irr fork we can force using it |
21:57 |
sapier |
I added a check and in case of error try to redownload the texture in GL_RGBA format |
21:57 |
nrzkt |
look at STK fork before wanting to forking irrlicht please |
21:57 |
est31 |
RealBadAngel, there is no point in doing a whole FORK when you can just have a patch like now |
21:58 |
sapier |
if this fails to ... well it'd not be worse then before ;-) |
21:58 |
nrzkt |
and look at irrlicht 1.9 preview changelog which is so huge |
21:58 |
RealBadAngel |
est31, lookin at instant crashes report is not any good too |
21:59 |
RealBadAngel |
better to have the reason in own hands, and dismiss it |
21:59 |
sapier |
imho a irrlicht fork would be a bad decision maintaining a 3d engine is at least as much work as minetest itself |
21:59 |
est31 |
RealBadAngel, that's the burden we have to bear as sole users of irrlicht |
21:59 |
est31 |
or well, perhaps there are other users too |
21:59 |
nrzkt |
sapier, i agree with you |
21:59 |
RealBadAngel |
est31, i suggest to see tux project |
22:00 |
sapier |
RealBadAngel: "in own hands" sounds like different taste of "not invented here"-syndrom ;-) |
22:00 |
est31 |
but on all distros I've checked, the only package using the irrlicht package is minetest |
22:00 |
RealBadAngel |
they have learned |
22:00 |
RealBadAngel |
we dont have to reinvent the wheel |
22:00 |
RealBadAngel |
we can choose to follow the right path |
22:00 |
sapier |
I'm gonna leave this discussion now as it's basically at same point it has been last year and the year before ;-) |
22:01 |
nrzkt |
RealBadAngel, i tested STK embedded into MT this weekend and this was not working out of the box :) |
22:01 |
VanessaE |
est31: "sole users of" is a bit of a scretch: http://irrlicht.sourceforge.net/project-platform/linux/ |
22:01 |
VanessaE |
(and MT isn't even listed there) |
22:01 |
est31 |
Error 503 Service Unavailable |
22:01 |
sapier |
unless there are at least two ppl having prooven to be capable of pushing a project in size of irrlicht and who say "yes I'm gonna fork irrlicht and fix all it's bugs" I won't change my opinion ;-) |
22:01 |
VanessaE |
O_o |
22:01 |
VanessaE |
reload. |
22:02 |
RealBadAngel |
nrzkt, yeah, together with me, you forgot to mention ;) |
22:02 |
est31 |
seems it got slashdotted VanessaE |
22:02 |
VanessaE |
well, there are ten projects listed on that page ()including SuperTuxKart) |
22:02 |
sapier |
and if those guys fork it I suggest calling it irrdark and we're just switching from irrlicht to that other project providing binary compatibility without it's bugs ;) |
22:03 |
VanessaE |
some tank battle game, some sort of puzzle game, couple other things, and a number of library/bindings packages. |
22:03 |
est31 |
but all not packaged in distros |
22:03 |
RealBadAngel |
and as one last nail, look at BAW |
22:03 |
VanessaE |
fair enough |
22:04 |
sapier |
I guess irrlicht is somehow outdated nowadays but at least I wont spend any time replacing it ... but if someone wants to replace it I'll encourage everyone to do it ... but don't expect others to help ;-) |
22:05 |
VanessaE |
sapier: zeno was working on that, but he left MT before finishing (or releasing) that code. :( |
22:05 |
RealBadAngel |
sapier, my point was tux and BAW are using irr too |
22:05 |
sapier |
I'd suggest talking to pilzadam he did at least a prototype replacement for irrlicht |
22:05 |
nrzkt |
sapier, irrlicht developpers are working a little bit these days, the project restarted |
22:05 |
sapier |
but I don't know how far his development ever progress |
22:05 |
nrzkt |
https://github.com/zaki/irrlicht/blob/master/changes.txt |
22:05 |
RealBadAngel |
we are not at the step we are starting to use an engine |
22:06 |
RealBadAngel |
we do know what we do want from the engine |
22:06 |
sapier |
long list of changes |
22:06 |
nrzkt |
this build works nearly with MT |
22:06 |
nrzkt |
i have crashes on some memcpy with specific textures, like the water texture, i'm trying to reproduce the MT code outside for the dev |
22:06 |
RealBadAngel |
we could at least make combined lightin system on top of bundled irr |
22:07 |
RealBadAngel |
if you all dont see most wanted reason |
22:07 |
sapier |
the interesting thing is android ogles branch is based on irrlicht 1.9 and doesn't seem to be any less stable then the previous version |
22:08 |
nrzkt |
sapier, you can try it, connect to my server (AppleTree) and tell me if your client crash due to irrlicht on water texture :) |
22:08 |
nrzkt |
i only tested master branch |
22:08 |
sapier |
nrzkt server name? |
22:09 |
nrzkt |
AppleTree in the serverlsit |
22:09 |
sapier |
server list on android ;-) |
22:09 |
nrzkt |
oh yes :D |
22:09 |
nrzkt |
pts.epixel-game.net 30000 |
22:10 |
sapier |
btw did anyone ever try to use the tree settings on android? |
22:10 |
nrzkt |
i think we should remove this android special menu to unify menus |
22:11 |
sapier |
nrzkt there's been a reason for the special menu, original menu wasn't to be changed even a single pixel and it just didn't fit on any screen of any android device I had |
22:11 |
sapier |
if those constraints have changed feel free to write a new menu ;-) |
22:12 |
sapier |
but be warned writing a full featured menu (spare time) takes about 3-6 months |
22:15 |
sapier |
strange I don't get any network connectivity with my current android client |
22:16 |
BlockMen |
can #3417 get merged? |
22:17 |
est31 |
if sb +1es it |
22:17 |
RealBadAngel |
water texture has separate shader |
22:17 |
RealBadAngel |
hardcoded |
22:17 |
RealBadAngel |
and not updated for a long time... |
22:17 |
BlockMen |
est31 : well, i do then |
22:17 |
nrzkt |
RealBadAngel, the bug wasn't on the shader, the crash was on the water block |
22:17 |
nrzkt |
when i open the inventory :) |
22:17 |
BlockMen |
will merge it then :P |
22:17 |
est31 |
ok |
22:18 |
est31 |
the thing PilzAdam commented about (including me) has been fixed |
22:18 |
est31 |
unless he commented about sth different than me |
22:18 |
est31 |
but I doubt that |
22:18 |
est31 |
if its proper for you, it can be merged |
22:19 |
est31 |
he had 13 days to point out the problem if he still has one with it :) |
22:21 |
sapier |
maybe we can use latest sdk ... could've tested wrong thing before |
22:22 |
nrzkt |
sapier, last release was done with the android sdk for android 5.0 if i remember |
22:23 |
BlockMen |
merged |
22:23 |
sapier |
I'm trying current version, I thought I'd need a system image for arm but that's completely wrong I do only need the sdk platform for android-10 ... which is still available |
22:24 |
BlockMen |
what about #3393? |
22:24 |
BlockMen |
hmmm doesnt seem to be interested |
22:26 |
sapier |
3393 seems to be right direction |
22:27 |
sapier |
I would've been glad if I did use a substruct when the automatic face movement property was added |
22:27 |
nrzkt |
BlockMen, you trigger a protocol break |
22:27 |
sapier |
ok recent sdk does work |
22:28 |
nrzkt |
the function is interesting BlockMen except the protocol break |
22:28 |
sapier |
where's a protocol break? |
22:28 |
BlockMen |
nrzkt : you mean https://github.com/minetest/minetest/pull/3393/files#diff-da1e28445fc732755d2b64683bd94ee4R858 ? |
22:28 |
BlockMen |
yes, needs to be changed back to longstring |
22:29 |
nrzkt |
yes :) |
22:29 |
sapier |
ohhh there you're removing things |
22:30 |
sapier |
thanks nrzkt I didn't see it |
22:34 |
BlockMen |
fixed |
22:36 |
nrzkt |
ty |
22:37 |
sapier |
https://github.com/minetest/minetest/pull/3454 anyone here to check android build? I didn't test oldarm mips and x86 by now as due to lack of build parallelity a single full build takes about an hour for me ... and I don't even have devices for thos archs |
22:38 |
nrzkt |
maybe i could test tomorrow but apk is not working on my phone |
22:38 |
BlockMen |
sapier: not doing android builds nor owning possible test device, sorry |
22:38 |
nrzkt |
also i have a gcc 5.3 on archlinux i don't know if it will be used |
22:38 |
est31 |
no, custom toolchain will be assembled for most things |
22:38 |
BlockMen |
can someone give here a second approve? #3358 |
22:39 |
sapier |
ci doesn't provide android build support does it? |
22:39 |
BlockMen |
last time im asking for such a simple thing |
22:39 |
nrzkt |
i had one |
22:39 |
nrzkt |
but i had a problem on it and no time to fix it then it's disable atm |
22:39 |
est31 |
sapier, travis CI has problems because of limited RAM. it crashes while building :) |
22:40 |
est31 |
sapier, so nrz provided a CI for us :) |
22:40 |
nrzkt |
only the rolling release is currently running |
22:40 |
nrzkt |
https://jenkins.unix-experience.fr/job/minetest-android-OfficialRepository-RollingRelease/ |
22:40 |
nrzkt |
https://jenkins.unix-experience.fr/job/minetest-Android-OfficialRepository/ |
22:40 |
nrzkt |
a problem on my java home i should fix :p |
22:41 |
sapier |
hmm maybe I should fix android makefile to not require manual changing it for switching between platforms |
22:42 |
sapier |
on the other hand if you mix up dependencies using different architectures I don't even wanna see the result |
22:43 |
|
Robert_Zenz joined #minetest-dev |
22:43 |
nrzkt |
sapier, if you have time to give a cmake build :p |
22:45 |
sapier |
for sure not |
22:45 |
nrzkt |
:p |
22:45 |
nrzkt |
i added a symlink for JAVA_HOME, we will see if the rolling release android iwll work |
22:45 |
sapier |
wrapping that huge load of duct tape in makefile format around of android build was more then enough pain |
22:46 |
BlockMen |
#3393 : protocol break fixed, travis passed, ready to merge |
22:49 |
RealBadAngel |
https://github.com/minetest/minetest/pull/3391 |
22:49 |
est31 |
one thing we should change IMO: the list of files should be autogenerated |
22:49 |
est31 |
perhaps it can be generated by CMAKE |
22:50 |
est31 |
then make calls cmake to generate the list or so |
22:50 |
sapier |
come on those android guys are joking ... if you tell them "use clang" but there ain't clang they just fallback to some compiler they find without even complaining |
22:54 |
nrzkt |
sapier, => https://jenkins.unix-experience.fr/job/minetest-android-OfficialRepository-RollingRelease/263/console |
22:54 |
nrzkt |
rolling release repaired |
22:54 |
nrzkt |
i re-enable the commit/PR builder |
22:54 |
sapier |
great :-) |
22:55 |
sapier |
can you add different build targets there? |
22:55 |
nrzkt |
i don't have hundreds of CPU :p |
22:55 |
nrzkt |
but if you need to build multiples target in one job i can |
22:55 |
sapier |
come one we do only have 4 build flavours atm ;-) |
22:56 |
nrzkt |
which target do you want exactly ? |
22:56 |
sapier |
mips and x86 ... I don't think we do need to test arm (non v7) |
22:56 |
nrzkt |
i cannot build for mips |
22:56 |
|
misprint joined #minetest-dev |
22:57 |
nrzkt |
x86 is a non sense because we have linux/gcc builder and it's same :) |
22:57 |
sapier |
but ndk does provide everything you need? |
22:57 |
est31 |
sapier, nrzkt one of you want to review https://github.com/minetest/minetest/pull/3393 |
22:57 |
sapier |
if you do have official ndk+sdk it'd be just a minor change to makefile ;-) |
22:57 |
nrzkt |
if i correctly understand we can have the nametag of every mob with this ? :p |
22:58 |
est31 |
yes |
22:58 |
nrzkt |
it's the official SDK/NDK, non updated on the jenkins since some time |
22:58 |
nrzkt |
but for the release build done on my pc it's up to date |
22:58 |
sapier |
sadly I couldn't use it for sure because it most likely conflicts the health bars I use |
22:58 |
sapier |
but that's not a reason not to merge it |
23:00 |
est31 |
the advantage of auto-generated lists would be that the android build bug reason #1 gets fixed |
23:00 |
est31 |
very often stuff works super fine on android |
23:00 |
nrzkt |
<3 http://pastie.org/10635006 thanks irrlicht 1.9 :D |
23:00 |
sapier |
what exactly do you mean with "auto-generated lists"= |
23:00 |
sapier |
? |
23:01 |
sapier |
est31 isn't there the actual error missing? |
23:02 |
est31 |
sapier, the error is that you forget to change the android list |
23:02 |
est31 |
you add file, you remove file you move file |
23:02 |
sapier |
ohh the file list :-) |
23:03 |
sapier |
not sure if there's enough benefit to really justify writing a source tree parser ... especially as those bugs are quite easy to fix |
23:04 |
nrzkt |
find src -name *.cpp |
23:04 |
nrzkt |
:p |
23:04 |
sapier |
nope |
23:04 |
est31 |
cmake has a whitelist as well |
23:04 |
est31 |
why not unify both :) |
23:05 |
sapier |
feel free to implement all the scripts within the current makefile as cmake extensions est31 ;-) |
23:06 |
sapier |
considering it still works after the original autor (me) didn't touch it for about a year I guess I have done something not tooo bad ;-) ... I'd have expected it to be completely broken |
23:06 |
sofar |
sadly, cmake makes portability harder, not easier |
23:06 |
sapier |
well android build is linux only |
23:07 |
est31 |
yeah, nrzkt tried to do it on freebsd |
23:07 |
est31 |
and failed :) |
23:07 |
nrzkt |
:p |
23:07 |
sapier |
but you're right cmake is almost a all or nothing build mechanizm either everything is cmake or it doesn't work |
23:08 |
nrzkt |
i didn't retry recently, linux emulation on FreeBSD works great :p |
23:08 |
sapier |
is there an android sdk for freebsd? |
23:08 |
BlockMen |
since no one claimed anything anymore im mergin #3393 in 10 minutes |
23:09 |
est31 |
BlockMen, find a better rule to justify |
23:09 |
est31 |
like to say "I am a core dev and I approve" est is one too, so we have two core devs who +1 it |
23:09 |
est31 |
:) |
23:11 |
BlockMen |
well, now you made it clear that you +1 it, before its was not clear for me |
23:11 |
sapier |
I wonder if I'll ever encounter a pull where ppl either not comment on any coding style issue or at least tell precise what they believe to be wrong |
23:11 |
BlockMen |
also, 2 more core devs around here not approving |
23:12 |
sapier |
BlockMen: could you rebase https://github.com/minetest/minetest/pull/3457 right after your merge? ;-) it's gonna be broken by it |
23:12 |
sapier |
but I'd expcet automatic rebase to be able to handle it ;-) |
23:13 |
RealBadAngel |
about my shaders PR, who have tried water surface shaders? |
23:13 |
BlockMen |
and im sick of this, really. then say: "not now", "i will look next time" or "i dont want to give it +1 for reason <xyz> |
23:13 |
est31 |
yeah I knew https://apkpure.com/pizzeria-craft-survival/com.touchapp.freddycraft |
23:13 |
BlockMen |
RealBadAngel, i will test your PR tomorrow. from look at it it seems fine to me |
23:13 |
est31 |
yet another rip off |
23:13 |
est31 |
but doesnt even point it out lol |
23:14 |
RealBadAngel |
BlockMen, this PR is just a gateway |
23:14 |
|
jin_xi joined #minetest-dev |
23:14 |
RealBadAngel |
alone it does nothing, absolutely nothing |
23:15 |
|
bleak_fire_ joined #minetest-dev |
23:15 |
BlockMen |
its obvious how to test its functionality ;) |
23:15 |
bleak_fire_ |
hi |
23:15 |
RealBadAngel |
so please do that |
23:15 |
RealBadAngel |
i do have 2 shaders rdy, water and lava |
23:15 |
BlockMen |
[00:13] BlockMen: RealBadAngel, i will test your PR tomorrow. |
23:15 |
sapier |
well I'd expect you to test the functionallity yourself BlockMen ;-) you wouldn't wanna merge it if it didn't do what you want ;-) |
23:16 |
RealBadAngel |
BlockMen, oke doke, waitin then :) |
23:16 |
sapier |
btw blockmen did you think about adding an "offset" parameter too? |
23:16 |
bleak_fire_ |
i dont know if anyone's been following me on #minetest |
23:16 |
bleak_fire_ |
but |
23:16 |
sapier |
e.g. if bounding box is way different then model size |
23:16 |
VanessaE |
wait, bleak_fire_ |
23:17 |
VanessaE |
until the current conversation is over. |
23:17 |
bleak_fire_ |
ok |
23:17 |
RealBadAngel |
btw, im damn proud of water flowing made with shaders, that piece took me a few weeks to polish |
23:17 |
BlockMen |
sapier, rule is 2 approves. and tested it with different models, was attached correct for me |
23:18 |
sapier |
I don't think it can be attached correct for all kinds of models but that's not topic of this pull so go ahead |
23:18 |
RealBadAngel |
i can say thats one of the best lookin effects ive ever made |
23:18 |
sofar |
RealBadAngel: shots? |
23:19 |
RealBadAngel |
no shots, you have to see it flowing |
23:19 |
sapier |
I've had that problem for my health bars in mobf which have different y offset dependent on mob model |
23:19 |
RealBadAngel |
its damn animation |
23:19 |
BlockMen |
RealBadAngel: btw shaders. could you get us better waving leaves shaders? they look awkward since paramats change |
23:19 |
sofar |
RealBadAngel: youtube? |
23:19 |
RealBadAngel |
BlockMen, im working on completely new waving models |
23:20 |
BlockMen |
RealBadAngel : +1 :) |
23:20 |
RealBadAngel |
please hold on with that |
23:20 |
sofar |
RealBadAngel: animaged gifv? :) |
23:20 |
RealBadAngel |
sofar, just grab the sources and try |
23:20 |
RealBadAngel |
im not CNN :P |
23:20 |
sofar |
I'm at work, lol |
23:20 |
sapier |
btw as you guys keep adding fancy stuff why didn't anyone ever add a non hacky sign ? |
23:21 |
sofar |
you mean signs_lib ? |
23:21 |
RealBadAngel |
sapier i did proper signs already |
23:21 |
sapier |
not an option |
23:21 |
RealBadAngel |
but im stopped with 3166 |
23:21 |
sapier |
signs_lib causes way to much texture load |
23:21 |
sofar |
I'm all for non-hacky signs |
23:21 |
sofar |
sign lag can be crazy, indeed |
23:21 |
RealBadAngel |
meta is holding textures generated on field change |
23:21 |
BlockMen |
still 103 open Prs :\ |
23:21 |
sapier |
we've seen this kill android clients years ago ;-) |
23:22 |
RealBadAngel |
but the engine is sending each meta change twice for some reason |
23:22 |
sofar |
client side rendering of sign text? |
23:22 |
sofar |
I know how to do that |
23:22 |
RealBadAngel |
its easy |
23:22 |
sofar |
just like custom tooltip text |
23:22 |
RealBadAngel |
but i need to get rid of those double updates |
23:22 |
sapier |
sofar: that's what I meant with "non hacky" |
23:22 |
RealBadAngel |
no, thats not such easy |
23:23 |
sapier |
and I could reuse that code for adding a laserpointer to 3d mode too ;-) |
23:23 |
RealBadAngel |
you have to generate texture on demand, and then flush it |
23:23 |
RealBadAngel |
runtime |
23:23 |
BlockMen |
RealBadAngel: what is with nrzkt 's comments, like https://github.com/minetest/minetest/pull/3166#discussion_r39203778 |
23:23 |
sofar |
a proper fishing rod would be nice |
23:23 |
RealBadAngel |
we doesnt do that yet at all |
23:23 |
sapier |
if I had to decide about fishing rod or signs ... welll ... that's hard .... I guess I'd take the signs ;-P |
23:24 |
nrzkt |
BlockMen, i added fixed on my forked minetest version but this PR still need the fixes if i remember |
23:24 |
RealBadAngel |
BlockMen, this is outdated, epixel is using the code for like 2 months already? |
23:24 |
nrzkt |
but this command is really great , for servers and for clients :) |
23:24 |
nrzkt |
i use a client compatible with it since 2 months and it's better for me |
23:24 |
BlockMen |
RealBadAngel : if you want it get merged you should update it? ;) |
23:25 |
RealBadAngel |
no, im asking for help with it |
23:25 |
RealBadAngel |
there should be a queue to get all the meta changes done in one tick |
23:25 |
RealBadAngel |
compress it (get rid of doubles in the process) and send them as one packet |
23:26 |
sapier |
sounds like something to be done out of main server loop |
23:26 |
RealBadAngel |
not really |
23:26 |
RealBadAngel |
theres already a loop to process all map edit events |
23:27 |
RealBadAngel |
but it was built to send or not send blocks |
23:27 |
RealBadAngel |
idk how to solve the situation |
23:27 |
sapier |
exactly it's not meant to do other things ;-) |
23:27 |
RealBadAngel |
but meta is completely other thing |
23:27 |
sapier |
it might work but you should look quite exactly at it |
23:28 |
RealBadAngel |
i think that this event should not be done there, completely |
23:28 |
RealBadAngel |
it shall sent block not sent to clients far away as before |
23:28 |
RealBadAngel |
or the older ones |
23:29 |
RealBadAngel |
but for new ones as described before |
23:30 |
sapier |
sorry don't know about the recent changes in there my knowledge about it might be updated or even missleading |
23:30 |
RealBadAngel |
the speedup is great, even without grouping meta changes |
23:30 |
RealBadAngel |
this is really a thing worth polishing |
23:30 |
sapier |
-updated + outdated |
23:30 |
est31 |
<RealBadAngel> this is really a thing worth polishing <----- full agree |
23:31 |
VanessaE |
agreed. |
23:31 |
sapier |
bleak_fire_: you did want to tell something? still there? |
23:31 |
RealBadAngel |
sapier, are you aware that moving shit in a chest is causing mesh upade for a block and all its neighbours? |
23:31 |
bleak_fire_ |
yes |
23:31 |
RealBadAngel |
this PR eliminates that |
23:31 |
sapier |
RealBadAngel: nope but that might explain the extreme delays |
23:31 |
RealBadAngel |
yes |
23:31 |
RealBadAngel |
with complex worlds |
23:31 |
bleak_fire_ |
i'm having pipeworks crashes |
23:32 |
BlockMen |
nrzkt : would you like to backport your fixed version from epixel to MT? |
23:32 |
bleak_fire_ |
usually when people are looking at/placing technic items |
23:32 |
bleak_fire_ |
"2015-12-15 23:09:46: ERROR[main]: Runtime error from mod 'pipeworks' in callback environment_Step(): /usr/share/minetest/builtin/game/item.lua:338: attempt to index local 'dropper' (a nil value)" |
23:32 |
VanessaE |
bleak_fire_: you mean you're having nil dropper reference crashes in buildin/item.lua |
23:32 |
sapier |
bleak_fire_: core related crashes? |
23:32 |
RealBadAngel |
for example, 8 solar arrays, one batbox, one swithichin box, and some cables below |
23:32 |
RealBadAngel |
all those nodes are causing circa 80 map edit events per tick |
23:32 |
nrzkt |
BlockMen, if i remember the fix :p |
23:32 |
RealBadAngel |
half of them doubled for no reason |
23:32 |
bleak_fire_ |
VanessaE suggested it might be another mod "dropping" something |
23:33 |
VanessaE |
sapier: my guess on bleak_fire_'s error is that the above referenced code just doesn't know how to deal with a nil 'dropper' as passed to it maybe by pipeworks or something. |
23:33 |
RealBadAngel |
now imagine whole damn VE's spawning area |
23:33 |
bleak_fire_ |
i dont understand what that means, and i've never used technic other than the protected chests. other people like it though |
23:33 |
RealBadAngel |
filled with machines |
23:33 |
sapier |
why is nil passed to it? |
23:33 |
VanessaE |
(and 'nil' player objects are undocumented in lua_api afaik) |
23:33 |
VanessaE |
beats me, I didn't write that code :P |
23:33 |
RealBadAngel |
each meta change causing mesh regen |
23:33 |
RealBadAngel |
clients are sitting duck, constantly regenerating meshes |
23:34 |
RealBadAngel |
because somebody called sending whole mapblock because of one meta change "A FIX" |
23:34 |
VanessaE |
RealBadAngel: ok ok we GET it. |
23:34 |
VanessaE |
merge your damn code already |
23:34 |
sapier |
bleak_fire_: which version are you using? |
23:34 |
nrzkt |
BlockMen, here are the commits related to this |
23:34 |
bleak_fire_ |
the latest |
23:35 |
VanessaE |
or someone else merge it for him, whatever |
23:35 |
bleak_fire_ |
at github |
23:35 |
bleak_fire_ |
usually every month or two i try and stay up to date |
23:35 |
RealBadAngel |
VanessaE, not until somebody will help or code with damn queue |
23:35 |
bleak_fire_ |
and report bugs as they happen |
23:35 |
sapier |
line 338 in item.lua doesn't contain "dropper"? |
23:35 |
nrzkt |
BlockMen, => http://pastie.org/private/do4yrvn3o3e3tthqiasjg |
23:35 |
RealBadAngel |
atm its half-done |
23:35 |
VanessaE |
well then will someone PLEASE HELP RealBadAngel finally? |
23:35 |
nrzkt |
the 4 commits related to this |
23:36 |
nrzkt |
first is the PR in its original case if i remember |
23:36 |
BlockMen |
sapier, not upstream version. after that secondary_on_use was added |
23:36 |
bleak_fire_ |
whoa is everyone pissed off at me? |
23:36 |
nrzkt |
others are fixed |
23:36 |
VanessaE |
good call, sapier |
23:36 |
VanessaE |
bleak_fire_: your minetest is outdated. |
23:36 |
sapier |
and for any usage of dropper in that file it's checked for nil before usage |
23:36 |
bleak_fire_ |
0.4.13? |
23:36 |
BlockMen |
nrzkt : thx |
23:36 |
VanessaE |
item.lua has that drop function at line 354, and there it IS protected. |
23:37 |
sapier |
I see you're using the official build bleak_fire_? |
23:37 |
VanessaE |
BlockMen: update to a dev build then |
23:37 |
BlockMen |
nrzkt : why is it necessary to register? |
23:37 |
VanessaE |
er, bleak_fire_ |
23:37 |
nrzkt |
because epixel is not ready for public usage |
23:37 |
bleak_fire_ |
http://slackbuilds.org/repository/14.1/games/minetest/ |
23:37 |
bleak_fire_ |
http://slackbuilds.org/repository/14.1/games/minetest_game/ |
23:37 |
VanessaE |
bleak_fire_: get it from the minetest repo and build it. |
23:37 |
nrzkt |
servers works very well, but there are missing features: p |
23:37 |
RealBadAngel |
bleak_fire_, if you are about to use rolling repo, please do update it on each commit |
23:38 |
nrzkt |
i will not sell your mail to others :) |
23:38 |
RealBadAngel |
not once a month or two :P |
23:38 |
VanessaE |
nrzkt: yeah, like not being able to use leveldb for the map :) |
23:38 |
nrzkt |
i use postgresql |
23:38 |
nrzkt |
it's better :p |
23:38 |
BlockMen |
nrzkt : do it, and give me 50% :P |
23:38 |
bleak_fire_ |
wait |
23:38 |
nrzkt |
BlockMen, lol |
23:39 |
bleak_fire_ |
so i should use the repo version of minetest, rather than the stable? |
23:39 |
BlockMen |
nrzkt : i dont get it, since ppl that want to look can do it, but meh, your code your rules |
23:39 |
bleak_fire_ |
do you mean github? |
23:39 |
nrzkt |
redis is a pain for huge map, we tested on appletree and minetest-france redis is not good for > 5 players connected on non huge memory servers |
23:39 |
sapier |
bleak_fire_: in your case yes |
23:39 |
VanessaE |
bleak_fire_: yes, github. |
23:39 |
bleak_fire_ |
ok |
23:39 |
RealBadAngel |
bleak_fire_, just compile it whenever you something has changed |
23:39 |
bleak_fire_ |
got it |
23:39 |
RealBadAngel |
+see |
23:39 |
nrzkt |
BlockMen, thanks :) |
23:40 |
bleak_fire_ |
will it work properly for users with stable versions? |
23:40 |
VanessaE |
RealBadAngel: well not on EVERY change... my client and server are both rather out of date now |
23:40 |
VanessaE |
bleak_fire_: should. |
23:40 |
RealBadAngel |
usually |
23:40 |
RealBadAngel |
but sometimes wont ;) |
23:40 |
BlockMen |
but i have to get some sleep... |
23:40 |
RealBadAngel |
its dev tree for christ sake ;) |
23:40 |
BlockMen |
gd night |
23:40 |
VanessaE |
(I don't actively track HEAD anymore. more like, just whenever it looks like I need to update) |
23:40 |
VanessaE |
night, BlockMen |
23:41 |
nrzkt |
idem good night all |
23:41 |
nrzkt |
i devel an interserver chatserver for those who are interested :p |
23:41 |
nrzkt |
i haven't the lua implementation by python implementation and C++ implementation is correct :p |
23:41 |
RealBadAngel |
im goin to take a nap too |
23:41 |
RealBadAngel |
tommorow i will polish 3d textures |
23:41 |
nrzkt |
it's very lightweight and portable :p |
23:41 |
RealBadAngel |
still lotsa work :) |
23:42 |
RealBadAngel |
but plants waving are GREAT |
23:43 |
RealBadAngel |
and im gonna kill paramat if he even starts to think loudly its not enough voxelish ;) |
23:44 |
VanessaE |
nonono it's PilzAdam who says that ;) |
23:45 |
RealBadAngel |
both dont like shaders |
23:45 |
RealBadAngel |
this is gonna be the first time i will restrict an effect to 16px only |
23:45 |
VanessaE |
well so much for HDX :P |
23:45 |
sapier |
gd night |
23:46 |
VanessaE |
night |
23:46 |
|
sapier left #minetest-dev |
23:46 |
RealBadAngel |
not really, there will be a way to texture it with higher resolution ones |
23:46 |
RealBadAngel |
but i mean there will be absolutely no reason to do so |
23:47 |
RealBadAngel |
default lookin pixel will be lookin good enough |
23:47 |
RealBadAngel |
its gonna be damn shaded cube on its own |
23:48 |
RealBadAngel |
atm wielded has one flaw, its flat |
23:49 |
RealBadAngel |
ive made it out of separate cubes |
23:49 |
RealBadAngel |
and those cubes are waving |
23:50 |
RealBadAngel |
so, there are 16 layers of waving micro nodes |
23:51 |
RealBadAngel |
damn plant is made out of single waving cubes, shaded and bumped |
23:52 |
VanessaE |
imho something like that is better as a proper surface mesh or something |
23:53 |
RealBadAngel |
i thought about using meshes for that |
23:53 |
RealBadAngel |
but better to generate them |
23:53 |
VanessaE |
it would be WAY fewer polys and it would work with any resolution of textures |
23:54 |
VanessaE |
keep plants flat, just split their faces up into 16x16 squares, and warp THOSE instead. |
23:54 |
RealBadAngel |
please wait until you see it live |
23:54 |
RealBadAngel |
i can just try to describe it, but words here are worth shit |
23:54 |
RealBadAngel |
same as with water shaders |
23:54 |
RealBadAngel |
you just have to see it |
23:55 |
VanessaE |
a youtube link will be sufficient for me :) |
23:55 |
RealBadAngel |
propably i will make a PR at the end of tommorow's session |
23:56 |
RealBadAngel |
or at least i will make a branch to test what ive done so far |
23:57 |
RealBadAngel |
hmmmm, btw, all those candies are possible only thx to tangent space commit |