| Time |
Nick |
Message |
| 02:35 |
|
Thermoriax joined #luanti |
| 02:36 |
|
crazylad joined #luanti |
| 03:49 |
|
MTDiscord joined #luanti |
| 04:42 |
|
aliasstilltaken joined #luanti |
| 05:00 |
|
MTDiscord joined #luanti |
| 05:37 |
|
apteryx_ joined #luanti |
| 06:23 |
|
cmc joined #luanti |
| 06:48 |
|
FeXoR joined #luanti |
| 07:58 |
|
swee joined #luanti |
| 08:37 |
|
mrcheese joined #luanti |
| 09:57 |
|
mrkubax10 joined #luanti |
| 10:20 |
|
jaca122 joined #luanti |
| 10:26 |
|
mrkubax10 joined #luanti |
| 12:01 |
|
Fleckenstein joined #luanti |
| 12:03 |
|
LizzyFleck joined #luanti |
| 12:27 |
|
PoochInquisitor joined #luanti |
| 12:51 |
|
gera joined #luanti |
| 13:41 |
|
serp3 joined #luanti |
| 14:00 |
|
turtleman joined #luanti |
| 14:01 |
|
user333_alt joined #luanti |
| 14:02 |
|
user333_alt left #luanti |
| 14:07 |
|
turtleman joined #luanti |
| 14:52 |
Bombo |
hi can someone help me fixing a bug in a mod? this is the issue: https://github.com/ElCeejo/creatura/issues/35 lines: https://github.com/ElCeejo/creatura/blob/main/mob_meta.lua#L212 i added this: 'if vel.x ~= vel.x then vel.x = 0 end' for x y and z right before self.object:set_velocity() and it seems to not crash anymore but not sure if that made sense |
| 14:55 |
Bombo |
adding a minetest.log shows it happens a lot that it is NaN |
| 14:57 |
[MatrxMT] |
<Blockhead256> it's kind of a band-aid solution. It fixes a NaN after it's happened, instead of at the cause. The stack trace might be able to tell you where it came form |
| 15:01 |
[MatrxMT] |
<Blockhead256> something's a bit suspicious about using the `and` keyword and the `or keyword` on floating point values in the first place. It's very rare in floating point to get an exact 0 figure |
| 15:02 |
[MatrxMT] |
<Blockhead256> well, that's used because `data` might have a falsy value. That sounds like a lack of initialization problem.. |
| 15:03 |
Bombo |
yes band aid, that's why i'm asking for help here ;) |
| 15:03 |
[MatrxMT] |
<birdlover32767> the 0 number is still true as bool? |
| 15:04 |
Bombo |
i don't know too much about lua, still learning |
| 15:05 |
[MatrxMT] |
<Blockhead256> `if 0.0 then print('boo') end` -> prints it. Yeah, 0.0 isn't falsy. But it seems more like it's there to guarantee init |
| 15:08 |
[MatrxMT] |
<Blockhead256> Bombo: the original reporter didn't have any more stack trace. Do you also not have any more stack trace? |
| 15:09 |
Bombo |
original reporter is me ;) |
| 15:09 |
|
kamdard__ joined #luanti |
| 15:09 |
[MatrxMT] |
<Blockhead256> oic |
| 15:09 |
Bombo |
no thats the only stack trace |
| 15:10 |
Bombo |
server shutting down, then these lnes show up |
| 15:13 |
Bombo |
i can add more minetest.log to check values |
| 15:15 |
[MatrxMT] |
<Blockhead256> how many birds in the box and how long? |
| 15:17 |
Bombo |
i spawned 10 but it really depends, right now it happens right when i join, yesterday i had to wait 30min |
| 15:17 |
MTDiscord |
<luatic> you need to trace the data flow back: where did the NaNs first arise? what populates self._movement_data? |
| 15:20 |
[MatrxMT] |
<Blockhead256> an owl came and ate my testing birds |
| 15:20 |
Bombo |
hnm if its not a number what is it, isn't there a way to find that out? or get an error msg why its a nan |
| 15:20 |
Bombo |
happens hehe |
| 15:22 |
[MatrxMT] |
<Blockhead256> any time you do floating point maths, there's a chance it comes out to NaN. Usually if you try to divide by zero. There's no immediate warning and no easy way to track it back to where it happened, it just turns everything it does arithmetic with into also NaNs until something handles it or crashes |
| 15:28 |
Bombo |
there were no nans for 15mins, all the birds were eaten, then i spawned one only and it started again |
| 15:28 |
Bombo |
now i restarted the server, 10 birds, non nans hmm |
| 15:32 |
[MatrxMT] |
<Blockhead256> creatura/boids.lua function creatura.get_boid_dir(self) |
| 15:33 |
[MatrxMT] |
<Blockhead256> there's a division by lift_no and at the top of the function it's initialized to 0. So there's the chance it stays that way if the ipairs() loop never runs. Don't know the odds of that |
| 15:35 |
[MatrxMT] |
<Blockhead256> "boids" is a flock simulation that's common in programming by the way. It would be to get the songbirds to fly as a flock together |
| 15:37 |
|
ireallyhateirc joined #luanti |
| 15:58 |
|
bgstack15 joined #luanti |
| 16:01 |
Bombo |
hm i get 'attempt to concatenate vert_vel (a nil value)' when i tried to minetest.log with "vertvel is: "..vert_vel or "nothing".. |
| 16:02 |
Bombo |
shouldnt that print nothing if vert_vel is nil? |
| 16:04 |
calculon |
looks like it needs parenthesis: "vertvel is: "..(vert_vel or "nothing").. |
| 16:15 |
|
jaca122 joined #luanti |
| 16:56 |
Bombo |
ah nice, well its nothing |
| 17:16 |
calculon |
btw "if its not a number what is it" -> funnily not a number is a number :) |
| 17:16 |
calculon |
ie it's a value with type "number" |
| 17:17 |
calculon |
and you can check if a value is nan with if x ~= x (because nan is never equal to another nan) |
| 18:56 |
|
qqq joined #luanti |
| 19:03 |
|
jaca122 joined #luanti |
| 19:05 |
|
jaca122 joined #luanti |
| 19:42 |
|
crazylad joined #luanti |
| 19:42 |
pgimeno |
you only get NaN for 0/0, not for other values /0, for those you get infinity |
| 19:49 |
Krock |
how about -NaN? |
| 19:50 |
pgimeno |
I'm not sure if Lua produces -NaN, that's indeterminate as opposed to not a number, but it's very platform-dependent |
| 19:50 |
user333_ |
player:set_pos({x=nan, y=nan, z=nan}) |
| 19:51 |
pgimeno |
0/0 is -nan; sqrt(-1) is nan |
| 19:51 |
Krock |
so nan * nan = -1? |
| 19:51 |
Krock |
unfortunately it propagates |
| 19:51 |
pgimeno |
no :) nan * nan = nan |
| 19:51 |
pgimeno |
yeah |
| 19:52 |
user333_ |
nan * nil |
| 19:55 |
pgimeno |
$ luajit -> =0/0 -> nan |
| 19:55 |
pgimeno |
$ lua -> =0/0 -> -nan |
| 19:57 |
pgimeno |
actually in puc lua, math.sqrt(-1) produces -nan in my system, but math.acos(2) produces nan |
| 20:02 |
|
PoochInquisitor joined #luanti |
| 20:09 |
|
mrkubax10 joined #luanti |
| 20:52 |
MTDiscord |
<sheriff_u3> Which is generally preferred: Using an item/node full name or alias in a crafting recipe. |
| 20:52 |
MTDiscord |
<sheriff_u3> Both work, just wanted to know if there was a standard. |
| 20:57 |
|
chilledfrogs joined #luanti |
| 23:34 |
|
panwolfram joined #luanti |
