| Time |
Nick |
Message |
| 00:44 |
|
deltib joined #minetest-dev |
| 01:09 |
|
Tablet_One joined #minetest-dev |
| 01:47 |
|
leat1 joined #minetest-dev |
| 02:27 |
|
Zeno` joined #minetest-dev |
| 02:44 |
|
Wayward_One joined #minetest-dev |
| 03:59 |
|
paramat joined #minetest-dev |
| 04:07 |
|
Miner_48er joined #minetest-dev |
| 04:29 |
|
Tablet_One joined #minetest-dev |
| 04:43 |
|
Hunterz joined #minetest-dev |
| 05:43 |
Zeno` |
What's the correct way in C++ to get the current mod directory? |
| 05:43 |
Zeno` |
asking because of #2481 |
| 05:43 |
ShadowBot |
https://github.com/minetest/minetest/issues/2481 -- having trouble registering schematic decorations |
| 05:55 |
|
leat1 joined #minetest-dev |
| 05:58 |
Zeno` |
maybe there is no way, currently |
| 05:58 |
Zeno` |
hm |
| 06:10 |
paramat |
hi hmmmm, here's my PR for fixing dust by dropping it from full_node_max.Y #2512 |
| 06:10 |
ShadowBot |
https://github.com/minetest/minetest/issues/2512 -- Mgv5/mgv7: Sprinkle dust from full_node_max.Y if chunk above is generated by paramat |
| 06:14 |
paramat |
ah just saw your comment zeno, interesting |
| 06:15 |
est31 |
Zeno`, you mean when loading the mod, or when its executing? |
| 06:16 |
est31 |
because the mod directory is only known when the mod is loaded |
| 06:17 |
est31 |
(with "loaded" I mean the "first run" through the file, that registers all functions) |
| 06:17 |
Zeno` |
well, it's when it's loading but not really |
| 06:17 |
Zeno` |
so there needs to be a mechanism added for prepending the mod path to the schematic filename |
| 06:18 |
est31 |
ShadowNinja already had that issue with whitelisting a mod's own directory |
| 06:18 |
est31 |
and we didnt find an easy solution |
| 06:18 |
est31 |
so only way would be to extend the api in some way |
| 06:19 |
est31 |
or to search all schematic dirs |
| 06:19 |
Zeno` |
ok well loading schematics as part of mapgen will have to stay broken until a solution is worked out |
| 06:19 |
est31 |
(and put that into the doc) |
| 06:19 |
Zeno` |
how will it know which world/ dir to look in? |
| 06:19 |
est31 |
thats what textures are doing |
| 06:20 |
est31 |
getting the world dir is possible |
| 06:20 |
est31 |
but no idea how, I would have to read the code, just like you can, but I guess faster :) |
| 06:20 |
Zeno` |
I'm not sure it should be in worlds/schems |
| 06:21 |
Zeno` |
Well, it certainly shouldn't be in there |
| 06:21 |
est31 |
why |
| 06:21 |
Zeno` |
Anyone with world edit can overwrite the mapgen schematic |
| 06:21 |
est31 |
for a mod? |
| 06:21 |
est31 |
yes then I agree it may not be in there |
| 06:21 |
Zeno` |
if it's in worlds/schems/ yeah.. so they should be in the mod directory |
| 06:22 |
est31 |
yes |
| 06:22 |
est31 |
I mean these are two different schematic uses here |
| 06:22 |
est31 |
worldedit uses it for, umm editing the world, as power user tool |
| 06:23 |
est31 |
and mapgen uses it as part of the "program" the user runs |
| 06:24 |
est31 |
so probably the easiest way is loading all schematic files at startup (from all mod dirs) and make it a best practice to prepend "modname_" before the schematic file name |
| 06:24 |
est31 |
perhaps even best to check that requirement |
| 06:24 |
est31 |
with "loading" I mean adding them to a vector. |
| 06:25 |
Zeno` |
what if different mods have a schematic with the same name? |
| 06:25 |
Zeno` |
same filename* |
| 06:26 |
Zeno` |
sorry, missed your comment |
| 06:28 |
hmmmm |
Zeno`, what's going on with that schematic loading issue?? |
| 06:28 |
hmmmm |
if I recall correctly, the schematic loading differentiates between the header sanity check and file not found |
| 06:28 |
Zeno` |
hmmmm, there is no path so it's not opening the file |
| 06:28 |
hmmmm |
oh |
| 06:29 |
hmmmm |
I need to fix that |
| 06:29 |
hmmmm |
hrmm, I wonder why this is all of a sudden broken |
| 06:29 |
Zeno` |
in Schematic::loadSchematicFromFile() I don't see much error checking at all :) |
| 06:29 |
hmmmm |
[02:17 AM] <Zeno`> so there needs to be a mechanism added for prepending the mod path to the schematic filename |
| 06:29 |
hmmmm |
I actually asked this... |
| 06:29 |
hmmmm |
there's a lua API to get the path of the current mod |
| 06:30 |
hmmmm |
modders use this api to construct the absolute filepath |
| 06:30 |
Zeno` |
I don't think that can be called from C++ though |
| 06:31 |
est31 |
I don't think its available after the loading step |
| 06:31 |
hmmmm |
https://forum.minetest.net/viewtopic.php?f=5&t=10901 |
| 06:31 |
hmmmm |
8 people thought it was okay |
| 06:31 |
paramat |
months ago i had this bug but just with the 'invalid schematic' error: implying a signature problem |
| 06:31 |
Zeno` |
how can it be ok? |
| 06:31 |
hmmmm |
i don't know actually |
| 06:31 |
hmmmm |
modders just think it's okay |
| 06:32 |
hmmmm |
this is why i asked |
| 06:32 |
Zeno` |
I can run minetest from any directory I want and it won't find the .mts file |
| 06:32 |
hmmmm |
so hold on |
| 06:32 |
hmmmm |
why was this broken |
| 06:32 |
hmmmm |
according to those people this only broke recently |
| 06:32 |
Zeno` |
I generally don't cd to the mod directory and ~/minetest/bin/minetest :) |
| 06:32 |
Zeno` |
I don't know when or how it was broken |
| 06:33 |
Zeno` |
I was just looking through the issues and decided that would be an easy one to trace (and it was) |
| 06:33 |
paramat |
heh oops i voted it was okay |
| 06:33 |
hmmmm |
well I'd rather figure that out because solving the problem directly will no doubt break reverse compatibility |
| 06:34 |
Zeno` |
Maybe the docs should specify that when registering the schematic they should prepend getModPath() (or whatever it's called) to the filename of the .mts |
| 06:34 |
Zeno` |
well, not should... must |
| 06:34 |
Zeno` |
although that's more of a workaround |
| 06:35 |
hmmmm |
like i said... I think that's how most people already do it |
| 06:35 |
hmmmm |
I don't know though |
| 06:35 |
hmmmm |
I never saw anybody's schematics code tbh |
| 06:35 |
Zeno` |
Obviously none of these people were aware of it https://github.com/minetest/minetest/issues/2481 :) |
| 06:36 |
est31 |
there isn't a function to get the path of the current mod, but there is one to get one by the mod name |
| 06:36 |
est31 |
minetest.get_modpath |
| 06:36 |
hmmmm |
i refuse to believe that many people didn't know that schematics were effectively broken until 6 days ago |
| 06:36 |
est31 |
(just for the record) |
| 06:36 |
hmmmm |
either i'm crazy or something weird is going on |
| 06:36 |
Zeno` |
est31, yeah so it could be ... schematic = minetest.getmodpath().."tree.mts", |
| 06:37 |
est31 |
minetest.getmodpath("coolmod") |
| 06:37 |
est31 |
that method needs a param read my comment |
| 06:37 |
Zeno` |
if that's how it works. Pretty ugly though |
| 06:37 |
est31 |
but yes |
| 06:37 |
hmmmm |
wait |
| 06:37 |
hmmmm |
something happened with sanitizing mod names |
| 06:37 |
hmmmm |
does it work if that fix is reverted? |
| 06:38 |
Zeno` |
which commit? |
| 06:38 |
hmmmm |
uhhh it was right before 0.4.12 went out |
| 06:38 |
hmmmm |
b0df67d |
| 06:38 |
Zeno` |
hmm, ok. I'll look but later. It's my 3 hour break time ;) |
| 06:38 |
Zeno` |
feel free to fix it while I'm gone hehe |
| 06:39 |
hmmmm |
no wait |
| 06:39 |
hmmmm |
that can't be it |
| 06:39 |
est31 |
~tell Zeno` because, after the loading step, we don't know the mod's path anymore. everything else (like taking the first element from the callstack) would also be bad. imagine a mod that acts as a library and calls that schematic function. |
| 06:39 |
ShadowBot |
est31: O.K. |
| 06:41 |
hmmmm |
yep I dunno |
| 06:41 |
est31 |
ow it was me? |
| 06:41 |
hmmmm |
maybe somebody can clue me in on how decorations were registered and worked before |
| 06:41 |
hmmmm |
and suddenly something broke |
| 06:42 |
est31 |
hmmmm, that change is only inside the main menu. If somebody tries to load a world with bad names, it already failed before |
| 06:42 |
hmmmm |
paramat: oh man the dust sprinkling is soo ugly |
| 06:43 |
hmmmm |
look at that |
| 06:43 |
paramat |
lol i thought so |
| 06:43 |
hmmmm |
it's not that you can help it |
| 06:43 |
|
Hunterz joined #minetest-dev |
| 06:43 |
hmmmm |
we need to come up with a smarter way to do this |
| 06:44 |
est31 |
do it like textures. |
| 06:44 |
hmmmm |
as it exists, there are too many issues with generating discrete structures that pop into a bordering mapblock |
| 06:44 |
hmmmm |
what do you mean 'do it like textures' |
| 06:45 |
est31 |
ah ok, you are already at another issue |
| 06:45 |
|
selat joined #minetest-dev |
| 06:45 |
est31 |
sorry, I thought you were reffering to the path problem |
| 06:45 |
paramat |
my PR there is just a first draft that works, but the implementation may be messy |
| 06:45 |
hmmmm |
hum i don't know how the texture path works |
| 06:45 |
|
nrzkt joined #minetest-dev |
| 06:45 |
est31 |
I'm not a mapgen person ;/ |
| 06:45 |
hmmmm |
paramat: it works in all cases? |
| 06:46 |
paramat |
yes apart from glitch at y = 47 |
| 06:46 |
hmmmm |
:) ofc |
| 06:46 |
paramat |
see description |
| 06:46 |
paramat |
yeah hehe |
| 06:46 |
nrzkt |
hi, i tested the damage fix i mentionned yesterday, it restore properly all the damage taken for many sources |
| 06:46 |
paramat |
concerning schematics, there is also this issue https://github.com/minetest/minetest/issues/2023 |
| 06:47 |
nrzkt |
without TNT damage me when enable_damage == false, with it works properly |
| 06:47 |
hmmmm |
paramat: i think that bug was fixed already. |
| 06:47 |
nrzkt |
i will push this fix now: https://github.com/nerzhul/minetest/commit/d1fa13c83ea855a9c76817e56fd8a5dbf0045246 |
| 06:47 |
paramat |
okay |
| 06:48 |
|
Anchakor_ joined #minetest-dev |
| 06:48 |
hmmmm |
really wish tenplus1 didn't disappear like that |
| 06:48 |
nrzkt |
VanessaE: i must take my train, please update you tree with current master for fixing your problems :) |
| 06:52 |
hmmmm |
paramat: LGTM, +2 |
| 06:54 |
paramat |
i like my method but the implementation is rough |
| 06:55 |
hmmmm |
just curious, does anybody have a use for generating a number from a PseudoRandom that fits a normal distribution with non-standard mean and variance? |
| 07:01 |
paramat |
not me |
| 07:02 |
hmmmm |
i was thinking like maybe you know how certain animals have more variability in intelligence? |
| 07:02 |
hmmmm |
what if you could have one particular type of mob that has a different mean but a much greater stddev |
| 07:03 |
hmmmm |
and thne another could have a very small stddev so it's very rare to find a mob of that type with that attribute |
| 07:03 |
hmmmm |
with an attribute that's considerably higher or lower, i mean |
| 07:04 |
hmmmm |
I guess this could be used for anything... HP, probability of finding items in a chest, probability of dropping a certain item when mining a special ore |
| 07:04 |
hmmmm |
right now I have a trivial method of estimating a discrete standard normal distribution given some min/max |
| 07:05 |
hmmmm |
but in order to get a normal distribution with a standard deviation != 1, i need to use something like the box-mueller transform |
| 07:07 |
paramat |
ah just saw your comment, i'll push the dust PR then for the short term, and try to improve it later |
| 07:08 |
paramat |
might work on it a bit more before pushing.. |
| 07:15 |
paramat |
so your 'generate cutoffs' method is no better? |
| 07:15 |
hmmmm |
=/ |
| 07:15 |
hmmmm |
i had an idea |
| 07:16 |
hmmmm |
it involves having 2 stages of mapblock generation and 3 distinct states |
| 07:16 |
hmmmm |
there'd be an in-between state that's like semi-there |
| 07:16 |
hmmmm |
it all made sense when I came up with it, but I forgot the finer details |
| 07:17 |
paramat |
i see. any reason why ravines are disabled in cavegen? |
| 07:17 |
hmmmm |
I think I disabled them because they were too common |
| 07:18 |
paramat |
i think i remember them in early mgv7 |
| 07:18 |
hmmmm |
and besides, they're not needed anymore because v5's cave generation method naturally creates ravines |
| 07:18 |
hmmmm |
which end up looking better too, imho |
| 07:18 |
paramat |
okay i might tune the rarity and see if i like them |
| 07:18 |
hmmmm |
have a blast.. |
| 07:19 |
paramat |
for mgv7 only though, and rare =) |
| 07:22 |
hmmmm |
what do you think about using noise2d() for deco placement |
| 07:22 |
hmmmm |
that would give so much finer control over the distribution than what exists currently |
| 07:24 |
hmmmm |
the only reason why it currently uses the pseudorandom method is because it's so computationally cheap |
| 07:24 |
paramat |
so calculating for every node? |
| 07:24 |
hmmmm |
yeah |
| 07:25 |
hmmmm |
that's 6400 per chunk but it's not that bad |
| 07:25 |
paramat |
the divlen method seems good though to reduce noise calculations |
| 07:25 |
hmmmm |
but look at what happens |
| 07:25 |
hmmmm |
if the rarity gets low enough it turns into 1 deco per division |
| 07:25 |
hmmmm |
and it ends up looking stupid |
| 07:26 |
paramat |
yeah i see |
| 07:26 |
hmmmm |
it's a little bit too discrete |
| 07:26 |
paramat |
possibly a good idea.. |
| 07:26 |
hmmmm |
if the calculated divlen is 2 then it'll place exactly 2 decos in that given area |
| 07:26 |
hmmmm |
that does not look random |
| 07:28 |
paramat |
the simplification might couteract calculating for every node |
| 07:29 |
paramat |
i mean in terms of processing load, your idea may even be faster |
| 07:29 |
hmmmm |
oh definitely not |
| 07:29 |
hmmmm |
nothing is going to beat the current placement method for speed |
| 07:30 |
hmmmm |
like..... i hate to say it but.. add a flag? :/ |
| 07:30 |
hmmmm |
"making superfluous config options is an excuse for poor engineering" |
| 07:31 |
paramat |
i'll think on this =) |
| 07:31 |
hmmmm |
the thing is that the pseudorandom division method works fine for decorations with a relatively high distribution |
| 07:32 |
hmmmm |
when we discrete-ize the perlin noise into a "number of decorations to place" value, that's the thing that kills it i think |
| 07:33 |
hmmmm |
what if we deal with fractions down to a certain resolution by increasing/decreasing the number of decos to place by another pseudorandom choice |
| 07:41 |
paramat |
an issue i'm having with schem decos is that even with a high fill ratio, for a dense jungle, there are patchy large areas missing trees, especially on the hills, it's very mysterious |
| 07:41 |
paramat |
but the grasses are dense everywhere, they work |
| 07:42 |
hmmmm |
ohh |
| 07:42 |
paramat |
this can be seen with https://github.com/paramat/biomesdev |
| 07:42 |
hmmmm |
either the documentation is inaccurate or you misunderstand |
| 07:43 |
paramat |
also, some of the large cacti schems are sunk into the ground in places |
| 07:43 |
hmmmm |
fill_ratio is the uniform ratio of the area to fill with decorations, ONLY USED IF there is no noise specified |
| 07:43 |
paramat |
yeah i know that =) |
| 07:43 |
hmmmm |
so if you have a noise_params specified, fill_ratio doesn't matter at all =] |
| 07:43 |
hmmmm |
then wut |
| 07:43 |
paramat |
it's either fill ratio or noiseparams |
| 07:44 |
paramat |
wut is the perfect term |
| 07:44 |
paramat |
anywau i'll make an issue |
| 07:44 |
paramat |
(anyway) |
| 07:45 |
paramat |
it's almost as if mountain terrain makes schem decos much less dense |
| 07:45 |
hmmmm |
ahh jungle trees i see you do use fill_ratio |
| 07:45 |
hmmmm |
aaaaahhh |
| 07:45 |
hmmmm |
you restrict it to biomes |
| 07:46 |
paramat |
yea but patchy within jungle biome |
| 07:46 |
hmmmm |
I was wondering :) I seriously hoped you didn't have dense jungles all over your map |
| 07:46 |
|
blaze joined #minetest-dev |
| 07:46 |
paramat |
and jungle grass is super dense everywhere as it should be |
| 07:46 |
hmmmm |
well |
| 07:47 |
hmmmm |
there's literally no difference between how the two are placed |
| 07:47 |
hmmmm |
i mean you see the code yourself |
| 07:47 |
paramat |
try biomesdev in mgv7/mgv5 if you want to see |
| 07:47 |
paramat |
yeah i read the schem placing code |
| 07:47 |
hmmmm |
is it patchy in a regular pattern? |
| 07:47 |
paramat |
nope, not related to altitude, block edges |
| 07:47 |
paramat |
etc. |
| 07:48 |
hmmmm |
it could just be the result of a crappy rand() implementation |
| 07:48 |
hmmmm |
yeah I'll try biomesdev |
| 07:48 |
paramat |
but then junglegrass works fine |
| 07:49 |
paramat |
you might find some sunken cacti schems too |
| 07:49 |
paramat |
seems to be an issue with schems not the placing code, at a guess |
| 07:50 |
paramat |
apart from that, mgv7 is looking beautiful, even before i tune the noiseparams, it will be awesome |
| 07:51 |
paramat |
i'll add in the mgv6 type cliffs, and make more use of the 3D noise terrain, try to combine the best of mgv5 and mgv6 |
| 07:53 |
paramat |
try it in mgv7, it's easier to see the problem.. affects all the trees but more noticeable in jungles because jungles have a uniform very high density of trees |
| 07:53 |
hmmmm |
what the f.. |
| 07:53 |
paramat |
often it seems any hills or mountain terrain makes tree density much lower |
| 07:53 |
hmmmm |
erm, I can't at the moment |
| 07:53 |
hmmmm |
the network code is crashing me |
| 07:53 |
paramat |
no prob |
| 07:55 |
paramat |
anyway must go |
| 07:55 |
|
paramat left #minetest-dev |
| 07:55 |
hmmmm |
i've gotta go too |
| 07:55 |
hmmmm |
it's 4 am |
| 07:56 |
hmmmm |
this is going to have to be a tomorrow thing |
| 08:07 |
|
Dartmouth joined #minetest-dev |
| 08:20 |
|
Calinou joined #minetest-dev |
| 08:27 |
|
kilbith joined #minetest-dev |
| 08:40 |
|
nrzkt joined #minetest-dev |
| 09:11 |
est31 |
wow app has > 100 users |
| 09:12 |
nrzkt |
i don't look today |
| 09:13 |
nrzkt |
77 users on thursday |
| 09:13 |
nrzkt |
tuesday, sorry :) |
| 09:14 |
est31 |
play store sais 100-500 |
| 09:14 |
est31 |
so >= 100 :) |
| 09:14 |
nrzkt |
the play store console stats are 2 days in the past, maybe not the main page |
| 09:14 |
nrzkt |
77 users but 17 doesn't update their application. |
| 09:15 |
nrzkt |
I will do another release for android this week-end to add some mods to minetest_game. |
| 09:16 |
nrzkt |
a problem reported by a user: The only problems are that the settings are not pointed to at the beginning with some kind of notification, and that the website says that this version does not exist. |
| 09:18 |
nrzkt |
can someone update our website ? |
| 09:21 |
|
paramat joined #minetest-dev |
| 09:22 |
paramat |
will push #2512 when checks are done |
| 09:22 |
ShadowBot |
https://github.com/minetest/minetest/issues/2512 -- Mgv5/mgv7: Sprinkle dust from full_node_max.Y if chunk above is generated by paramat |
| 09:25 |
nrzkt |
okay :). Paramat, can you answer to Zeno on #2429 for making this PR mergeable by him ? |
| 09:25 |
ShadowBot |
https://github.com/minetest/minetest/issues/2429 -- Update credits by Zeno- |
| 09:52 |
nrzkt |
on the forum we are talking about making an android_game for android, based on minetest_game adding some features for mobile users. What do you think about the creation of a android_game repository in our organization ? |
| 09:54 |
paramat |
okay |
| 09:56 |
|
MinetestForFun joined #minetest-dev |
| 10:03 |
paramat |
as for an android game repo, i'm neutral :] |
| 10:05 |
kilbith |
paramat, i've noticed some large zones of gravel spread over the desert stones recently, this is intentional ? |
| 10:05 |
kilbith |
example : https://lut.im/vAZEiGhP/T7E2bhsl |
| 10:06 |
paramat |
could be a rare gravel biome, are these common? are they square when seen from above? |
| 10:07 |
kilbith |
not common but not ultra-rare either |
| 10:07 |
paramat |
i'll check the code.. |
| 10:08 |
kilbith |
and no, not only squarred - the screenie above shows half the part of that zone |
| 10:08 |
paramat |
good |
| 10:09 |
kilbith |
the other part is accross the climb on the left |
| 10:10 |
paramat |
gravel biomes do occur in deserts, and occur when mud noise goes negative, which is rarely. so i think this is nothing to worry about, as long as they are fairly rare |
| 10:10 |
kilbith |
well, this is not looking bad anyhow |
| 10:11 |
kilbith |
and adds variety |
| 10:11 |
paramat |
these are like nyancats, reason for celebration |
| 10:14 |
paramat |
they are dependant on world seed, so certain seeds will create these near world centre |
| 10:15 |
paramat |
okay now pushing 2512.. |
| 10:16 |
paramat |
will take about 15min |
| 10:27 |
paramat |
done |
| 10:33 |
|
Dartmouth joined #minetest-dev |
| 10:35 |
|
est31 joined #minetest-dev |
| 10:36 |
|
AnotherBrick joined #minetest-dev |
| 10:36 |
|
Amaz joined #minetest-dev |
| 10:39 |
|
paramat left #minetest-dev |
| 10:48 |
celeron55 |
https://forum.minetest.net/viewtopic.php?p=172149#p172149 |
| 10:49 |
celeron55 |
i wonder if this could be helped in some website-organizational way or something |
| 10:50 |
celeron55 |
like, having a page that beginning modders are directed to, which is maintained by experienced modders and contains recommendations on what to use as bases for certain things |
| 10:50 |
celeron55 |
because... we don't have any of that |
| 10:52 |
celeron55 |
i guess it's doable in the developer wiki as long as people decide it's a useful thing to have |
| 10:55 |
|
Calinou_ joined #minetest-dev |
| 11:09 |
|
neoascetic joined #minetest-dev |
| 11:09 |
neoascetic |
Hi all. Is there a concept of client-side only mods? |
| 11:10 |
neoascetic |
For example, "ambience" mod - there is not need to control it by server since client may done all computations by its own |
| 11:10 |
est31 |
neoascetic, not yet, perhaps in future. |
| 11:11 |
neoascetic |
I hope this will be implemented soon |
| 11:11 |
est31 |
it has high demands on security |
| 11:11 |
neoascetic |
By what reason? |
| 11:12 |
est31 |
right now, mods on a server can in theory execute any code on that server. If you run minetest as root, they can nuke your system |
| 11:12 |
est31 |
it was tolerable for the server side, as you can give the server its own user |
| 11:13 |
est31 |
but client shouldn't be patched later on |
| 11:13 |
est31 |
it should be secure from the start |
| 11:13 |
est31 |
but thats not the blocker issue I think. |
| 11:13 |
neoascetic |
Probably we are talking about different things. |
| 11:14 |
neoascetic |
Client-side mods running only on client-side and has no access to server at all |
| 11:14 |
neoascetic |
I. e. their "server" is client |
| 11:14 |
neoascetic |
as a proxy |
| 11:15 |
est31 |
yes, but do you want some software that downloads other software onto your computer, and executes it? |
| 11:15 |
est31 |
btw. security is more or less done for the server side, and perhaps its suiting (as in secure enough) for client |
| 11:15 |
neoascetic |
Currently servers already downloads mods on your computer. Client-side mods are totally controlled by user only |
| 11:16 |
neoascetic |
As extra-plugins, server event may not know about them |
| 11:16 |
est31 |
tell me about use cases. |
| 11:16 |
est31 |
what would such a mod do? |
| 11:17 |
neoascetic |
Ambience mod. What the reason to track each user and decide which sound to play on server-side? It is pure client-side problem IMO |
| 11:17 |
neoascetic |
Music/sounds/etc |
| 11:18 |
Calinou |
there's no client-side API for playing sounds |
| 11:18 |
Calinou |
and even less client-side Lua |
| 11:18 |
neoascetic |
And mods may be executed in sandbox, as far as I know it is pretty easy in lua |
| 11:18 |
Calinou |
the problem of client-side Lua is similar to the JavaScript trap |
| 11:18 |
Calinou |
ie. server might download non-free code and run it without your consent |
| 11:18 |
Calinou |
this can be fixed with checking license headers, but people will have to add such headers |
| 11:18 |
Calinou |
(like LibreJS does) |
| 11:18 |
est31 |
ow please no fsf flame |
| 11:18 |
neoascetic |
Oh |
| 11:18 |
|
nrzkt joined #minetest-dev |
| 11:19 |
Calinou |
est31, it's a real problem :( |
| 11:19 |
Calinou |
here we can easily fix it |
| 11:19 |
neoascetic |
that's trap |
| 11:19 |
Calinou |
if client-side Lua is ever implemented, this is to be done, ideally officially |
| 11:19 |
est31 |
Calinou, so you are executing closed source microcode right now, and, are you hurt? |
| 11:19 |
Calinou |
est31, just shut up if you're flaming |
| 11:20 |
nrzkt |
celeron55 legitimate |
| 11:20 |
est31 |
ok Calinou you too please. |
| 11:20 |
neoascetic |
Anyway, I think it would be a great "feature" |
| 11:20 |
est31 |
the problem for client mods is that you get into the situation of having different versions or servers require them |
| 11:20 |
Calinou |
est31, else we might end up in Quake land where lots of mods have proprietary game logic running on your side… |
| 11:20 |
est31 |
and that isnt a situation you want to start with. |
| 11:20 |
neoascetic |
Put some execution on client-side (even from server-provided mods) and will improve client performance in some cases |
| 11:21 |
Calinou |
and there's absolutely nothing you can do against it, except not playing on these servers |
| 11:21 |
Calinou |
that game logic comes in .so format by the way |
| 11:21 |
Calinou |
so it's binary |
| 11:21 |
Calinou |
thankfully Lua is not compiled usually |
| 11:21 |
Calinou |
but it's still a problem nonetheless |
| 11:21 |
Calinou |
neoascetic, we'll talk about it once we have proper sandboxed Lua :p |
| 11:25 |
neoascetic |
Calinou what problems are with Lua sandboxing? |
| 11:25 |
Calinou |
none, we just need to do it |
| 11:25 |
est31 |
its done and the PR is "in labor" |
| 11:26 |
est31 |
aka short before merge |
| 11:26 |
Calinou |
as for license check: http://paste.ubuntu.com/10626866/ |
| 11:26 |
Calinou |
feel free to make it disabled by default, but I'd like such a thing anyway |
| 11:27 |
est31 |
neoascetic, here: https://github.com/ShadowNinja/minetest/commits/security-2 |
| 11:28 |
est31 |
PR is #1606 but outdated. |
| 11:28 |
ShadowBot |
https://github.com/minetest/minetest/issues/1606 -- Add mod security by ShadowNinja |
| 11:41 |
|
DFeniks joined #minetest-dev |
| 11:49 |
|
Hunterz joined #minetest-dev |
| 12:04 |
|
Zeno` joined #minetest-dev |
| 12:24 |
|
Calinou joined #minetest-dev |
| 12:36 |
|
Aaron1011 joined #minetest-dev |
| 12:38 |
|
MinetestForFun joined #minetest-dev |
| 12:44 |
|
neoascetic joined #minetest-dev |
| 12:47 |
Zeno` |
whoa, just reading the logs. |
| 12:47 |
Zeno` |
celeron55 says that buildcraft is a minecraft "ripoff"? |
| 12:48 |
Zeno` |
how? |
| 12:49 |
|
Aaron1011 joined #minetest-dev |
| 12:50 |
est31 |
no, he said its a mine*test* "ripoff". :P |
| 12:50 |
Zeno` |
is it though? |
| 12:51 |
neoascetic |
I am trying to fight agains #1614. Even small window resize cause all "click targets" move a few pixels above (as I may guess something about window bar size). Where I have to look at first? |
| 12:51 |
ShadowBot |
https://github.com/minetest/minetest/issues/1614 -- [OSX] Resizing window makes mouse input stop |
| 12:53 |
est31 |
https://en.wiktionary.org/wiki/rip-off#English |
| 12:53 |
est31 |
"A copy, especially one that is illegal or inferior." |
| 12:54 |
est31 |
I'm not sure about the illegal part, but for inferiority you can say that it doesn't use most recent engine and has ads |
| 12:54 |
est31 |
it has nicer controls though |
| 12:55 |
est31 |
perhaps they can be merged. |
| 12:55 |
* est31 |
wonders how the've solved the input problem |
| 12:55 |
est31 |
currently its solved bad ofc |
| 12:55 |
est31 |
(inside minetest master) |
| 12:56 |
est31 |
you tap on a text input field, it opens a native android one. no keyboard appears, you have to tap it again. |
| 12:57 |
est31 |
then you press on "go" on the keyboard, and you are back at the native input |
| 12:57 |
est31 |
then you have to figure out a way how to exit that one |
| 12:57 |
est31 |
and then you're back to formspec |
| 12:58 |
Zeno` |
hmm |
| 12:59 |
|
MinetestForFun joined #minetest-dev |
| 13:02 |
Zeno` |
the current android controls are terrible |
| 13:03 |
nrzkt |
+1 |
| 13:03 |
Zeno` |
I really have nfi how anyone can even play for more than 5 minutes hehe |
| 13:29 |
|
Tablet_One joined #minetest-dev |
| 13:31 |
|
DFeniks joined #minetest-dev |
| 13:32 |
|
Wayward_One joined #minetest-dev |
| 14:19 |
|
Wayward_One joined #minetest-dev |
| 14:38 |
kilbith |
est31, the PPA builds are still in Debug mode |
| 14:40 |
est31 |
ppa stable or unstable? |
| 14:40 |
kilbith |
daily-builds |
| 14:41 |
est31 |
yes thats the correct term |
| 14:41 |
kilbith |
dunno if the stable one is in Debug too |
| 14:41 |
est31 |
so why should they be no debug builds anymore? |
| 14:42 |
kilbith |
because they're slower than the Release one |
| 14:44 |
nrzkt |
but they permit to provide useful informations for debuging. For daily build Debug is better than release. Please keep Release for release. |
| 14:44 |
est31 |
yes thats true but aren't daily builds targeted for people who want to contribute to testing? |
| 14:44 |
nrzkt |
i think it's the goal est31 |
| 14:44 |
kilbith |
that's right |
| 14:44 |
nrzkt |
people who want stable builds use release, not daily |
| 14:48 |
kilbith |
should "minetestc55" package be suppressed on Launchpad ? |
| 14:49 |
kilbith |
it can lead to confusion |
| 14:51 |
VanessaE |
yes. |
| 14:52 |
VanessaE |
(sorry, c55 :) ) |
| 14:54 |
kilbith |
also that : https://lut.im/MyvIQZft/igDSTbDu |
| 14:54 |
kilbith |
to nuke |
| 14:56 |
VanessaE |
yeah, get rid of them since no one's updating them |
| 14:59 |
est31 |
good idea |
| 15:24 |
|
selat joined #minetest-dev |
| 15:26 |
est31 |
sfan5, you there? |
| 15:28 |
est31 |
sfan5 or any other german speaker: is #2490 ok ? |
| 15:28 |
ShadowBot |
https://github.com/minetest/minetest/issues/2490 -- Update and complete German translation and fix mistakes (second try) by Wuzzy2 |
| 15:28 |
est31 |
for me it is |
| 15:28 |
sfan5 |
est31: yes |
| 15:29 |
sfan5 |
in yes I'm here |
| 15:33 |
|
Calinou joined #minetest-dev |
| 15:34 |
|
Amaz joined #minetest-dev |
| 15:37 |
|
DFeniks joined #minetest-dev |
| 15:44 |
|
VargaD joined #minetest-dev |
| 15:45 |
est31 |
kilbith, you sure its also in the repo? |
| 15:45 |
est31 |
I cant find the package |
| 15:46 |
est31 |
https://launchpad.net/~minetestdevs/+archive/ubuntu/stable/+packages |
| 15:46 |
est31 |
https://launchpad.net/~minetestdevs/+archive/ubuntu/daily-builds/+packages |
| 16:26 |
nrzkt |
sfan5 if you can make a comment to approve #2490 on github we can push :) |
| 16:26 |
ShadowBot |
https://github.com/minetest/minetest/issues/2490 -- Update and complete German translation and fix mistakes (second try) by Wuzzy2 |
| 16:30 |
kilbith |
est31, on the c55 personal repo |
| 16:31 |
kilbith |
oops, too late |
| 16:34 |
|
fz72 joined #minetest-dev |
| 16:42 |
fz72 |
Zeno`: Any news on #2382? |
| 16:42 |
ShadowBot |
https://github.com/minetest/minetest/issues/2382 -- Remap Mouse and Key Buttons by fz72 |
| 16:52 |
|
MinetestForFun joined #minetest-dev |
| 16:58 |
|
Lunatrius joined #minetest-dev |
| 17:30 |
|
Hunterz joined #minetest-dev |
| 17:36 |
|
fz72 joined #minetest-dev |
| 17:43 |
|
VargaD joined #minetest-dev |
| 17:46 |
|
hmmmm joined #minetest-dev |
| 17:54 |
|
ElectronLibre joined #minetest-dev |
| 17:59 |
|
Robert_Zenz joined #minetest-dev |
| 18:02 |
|
luizrpgluiz joined #minetest-dev |
| 18:06 |
|
Krock joined #minetest-dev |
| 18:21 |
|
luizrpgluiz left #minetest-dev |
| 18:24 |
|
rubenwardy joined #minetest-dev |
| 18:26 |
|
est31 joined #minetest-dev |
| 18:29 |
est31 |
kilbith, so, which repo do you use exactly? |
| 18:29 |
est31 |
I don't see any PPAs directly assigned to c55, but perhaps I'm wrong |
| 18:29 |
kilbith |
ppa:minetestdevs/daily-builds |
| 18:30 |
est31 |
and that shows you those outdated packages? |
| 18:30 |
kilbith |
but even without it, you can always see 'minetestc55' |
| 18:30 |
* est31 |
adds it myself... |
| 18:33 |
kilbith |
https://code.launchpad.net/~minetestdevs/+archive/ubuntu/stable/+packages |
| 18:33 |
kilbith |
that package is in the repo |
| 18:34 |
est31 |
kilbith, which distro are you on? |
| 18:34 |
est31 |
also, which release? |
| 18:34 |
kilbith |
lubuntu 14.04 |
| 18:35 |
est31 |
ok, minetestc55 is there, but none of the other packages you referred to. |
| 18:36 |
est31 |
mhh I dont think c55 wants them to be removed |
| 18:36 |
kilbith |
why ? they're outdated |
| 18:36 |
est31 |
(as they are basically old versions of minetest) |
| 18:36 |
est31 |
perhaps make an archive ppa... |
| 18:36 |
kilbith |
they're not even stable, just middle in dev |
| 18:38 |
est31 |
? |
| 18:39 |
kilbith |
[old stable]--------[minetestc55]--------[new stable] |
| 18:39 |
kilbith |
clearer ? |
| 18:39 |
est31 |
yes I know what you mean |
| 18:39 |
est31 |
but what brings you to the assumption |
| 18:39 |
est31 |
the date? |
| 18:40 |
kilbith |
minetestc55 : 2013 11 17 |
| 18:41 |
kilbith |
which was 0.4.7-dev |
| 18:43 |
est31 |
btw are you referring to daily builds or to stable? |
| 18:43 |
kilbith |
daily-builds PPA |
| 18:46 |
est31 |
hm, as it seems only the latest versions are kept tehre |
| 18:46 |
est31 |
you are right, those should be removed. |
| 18:50 |
est31 |
(I was assuming you talked about the stable repo, sorry) |
| 18:51 |
Calinou |
by the way, Debian package has been updated to 0.4.12 recently |
| 18:51 |
Calinou |
sid users can enjoy 0.4.12 :) |
| 18:51 |
est31 |
yes, very nice :) |
| 18:51 |
est31 |
I guess it came too late for ubuntu 15.04 though |
| 18:52 |
kilbith |
yes, it's frozen right now |
| 18:52 |
Calinou |
so is jessie |
| 19:21 |
sfan5 |
uhh |
| 19:21 |
sfan5 |
the packages should be called "minetest" |
| 19:22 |
est31 |
I'll remove the minetestc55 packages from the daily ppa |
| 19:27 |
est31 |
kilbith, sfan5, the cleaned list: https://launchpad.net/~minetestdevs/+archive/ubuntu/daily-builds |
| 19:28 |
sfan5 |
looks better |
| 19:28 |
kilbith |
https://launchpad.net/~minetestdevs/+archive/ubuntu/stable |
| 19:29 |
kilbith |
^ clean-up needed |
| 19:29 |
kilbith |
oh, nvm |
| 19:29 |
sfan5 |
do we still support saucy & quantal? |
| 19:30 |
est31 |
both of them are EOLed |
| 19:30 |
est31 |
officially. |
| 19:30 |
kilbith |
10.04 ends up the next month afaik |
| 19:30 |
est31 |
server, yes desktop no, that ended in 2012 |
| 19:31 |
kilbith |
nobody uses it anyway |
| 19:31 |
est31 |
we don't offer precise packages |
| 19:32 |
est31 |
that was already since 0.4.11 though, and perhaps before too |
| 19:32 |
est31 |
(precise is currently still getting security updates) |
| 19:35 |
Calinou |
saucy is EOL, quantal is EOL |
| 19:35 |
Calinou |
10.04 server ends in april 30 (client has already ended) |
| 19:35 |
|
DFeniks_ joined #minetest-dev |
| 19:35 |
|
DFeniks joined #minetest-dev |
| 19:35 |
|
DFeniks__ joined #minetest-dev |
| 19:35 |
Calinou |
we should offer precise packages until 2017 |
| 19:35 |
est31 |
problem is that the build is failing |
| 19:36 |
|
DFeniks joined #minetest-dev |
| 19:37 |
est31 |
This for example: https://launchpad.net/~minetestdevs/+archive/ubuntu/stable/+build/7090191 |
| 19:37 |
est31 |
undefined reference to `snappy::RawCompress(char const*, unsigned int, char*, unsigned int*)' |
| 19:38 |
kilbith |
this is related with leveldb |
| 19:38 |
est31 |
hm, you suggest removing leveldb for precise? |
| 19:39 |
Calinou |
try it |
| 19:40 |
kilbith |
est31, look what i see when i remove the daily-builds PPA right now : https://lut.im/sZSZ79Af/65XBMCZs |
| 19:41 |
kilbith |
oops, sorry, forget to add the stable PPA |
| 19:42 |
kilbith |
although those outdated mods still needs to go |
| 19:42 |
sfan5 |
est31: leveldb is broken in precise, remove it |
| 19:42 |
est31 |
kilbith, can you do "apt-cache policy minetest-mod-worldedit"? |
| 19:42 |
Calinou |
I'm not sure if they would accept deletion requests of packages |
| 19:42 |
Calinou |
but they definitely look unmaintained by the original maintainer |
| 19:42 |
Calinou |
so someone either has to take them over, or remove them |
| 19:42 |
kilbith |
est31: nvm, just missed to add the PPA |
| 19:42 |
est31 |
Calinou, we first need to find out which PPA hosts them. |
| 19:43 |
Calinou |
https://packages.debian.org/sid/minetest-mod-moreblocks |
| 19:43 |
Calinou |
it's in Debian packages |
| 19:43 |
Calinou |
it's almost 2 years old now |
| 19:44 |
est31 |
sfan5, trying right now to disable leveldb |
| 19:44 |
Calinou |
can you disable it only for precise? |
| 19:44 |
est31 |
yes |
| 19:44 |
est31 |
thats my goal |
| 19:44 |
est31 |
it makes stuff harder though |
| 19:45 |
est31 |
I know of that hammer method we dont want, then a duplication method thats horribly ugly, and now I wonder whether there is another method... |
| 19:47 |
est31 |
sorry found a RL bug... |
| 19:59 |
|
devmarth joined #minetest-dev |
| 20:07 |
* est31 |
wonders how all that stuff worked in the past |
| 20:08 |
est31 |
two different changelogs |
| 20:43 |
|
Tablet_One joined #minetest-dev |
| 20:54 |
|
zat joined #minetest-dev |
| 20:56 |
|
zat joined #minetest-dev |
| 20:58 |
est31 |
Calinou, kilbith, sfan5: https://code.launchpad.net/~minetestdevs/+archive/ubuntu/stable/+recipebuild/891102 |
| 20:58 |
est31 |
build is triggered lets view what happens |
| 20:58 |
est31 |
(with leveldb disabled) |
| 21:06 |
|
Miner_48er joined #minetest-dev |
| 21:12 |
|
zat joined #minetest-dev |
| 21:14 |
|
ElectronLibre left #minetest-dev |
| 21:21 |
|
disablec1 joined #minetest-dev |
| 21:37 |
|
Calinou joined #minetest-dev |
| 21:40 |
|
Wayward_One joined #minetest-dev |
| 22:06 |
est31 |
so builds should work now |
| 22:06 |
VanessaE |
what's broken about leveldb in precise? |
| 22:06 |
est31 |
VanessaE, https://launchpad.net/~minetestdevs/+archive/ubuntu/stable/+build/7090191 |
| 22:07 |
VanessaE |
oh, broken like that eh |
| 22:08 |
VanessaE |
(misread the channel backscroll) |
| 22:24 |
|
leat1 joined #minetest-dev |
| 22:27 |
|
MinetestForFun joined #minetest-dev |
| 23:33 |
Sokomine |
does anyone know how that color filter for textures work? how to change a texture in color without having to add a new one? |
| 23:34 |
|
Tablet_One joined #minetest-dev |
