| Time |
Nick |
Message |
| 00:11 |
|
Jordach joined #minetest-dev |
| 00:20 |
|
Miner_48er joined #minetest-dev |
| 00:39 |
|
ANAND joined #minetest-dev |
| 00:48 |
|
erlehmann joined #minetest-dev |
| 00:56 |
|
ANAND joined #minetest-dev |
| 01:48 |
|
Taoki joined #minetest-dev |
| 03:31 |
oil_boi |
Now: should the new jump style be a settings override or should it be a clientside thing? |
| 03:41 |
|
reductum joined #minetest-dev |
| 04:35 |
|
calcul0n joined #minetest-dev |
| 05:33 |
|
erlehmann joined #minetest-dev |
| 05:58 |
|
ANAND joined #minetest-dev |
| 06:13 |
|
realzies joined #minetest-dev |
| 06:15 |
oil_boi |
#9980 8) |
| 06:15 |
ShadowBot |
https://github.com/minetest/minetest/issues/9980 -- Jumping fix by oilboi |
| 06:15 |
oil_boi |
So I decided on physics override, so people can join servers and have a choice on the server welcome screen |
| 06:16 |
oil_boi |
"Using a mod that does not exist yet" |
| 07:56 |
|
Beton joined #minetest-dev |
| 08:00 |
|
ShadowNinja joined #minetest-dev |
| 08:08 |
|
oil_boi joined #minetest-dev |
| 08:11 |
|
Darcidride joined #minetest-dev |
| 08:59 |
oil_boi |
A note: I just fixed the jitter glitch when jumping up over a node |
| 09:11 |
oil_boi |
#9983 |
| 09:11 |
ShadowBot |
https://github.com/minetest/minetest/issues/9983 -- Stop stair detection from jolting up player and entities by oilboi |
| 09:11 |
oil_boi |
That bug was so annoying |
| 09:30 |
|
oil_boi joined #minetest-dev |
| 09:47 |
sfan5 |
merging game#2695 in 5 minutes |
| 09:47 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/2695 -- Use node particles for TNT explosions by sfan5 |
| 10:08 |
|
oil_boi joined #minetest-dev |
| 10:27 |
|
NetherEran joined #minetest-dev |
| 10:31 |
|
proller joined #minetest-dev |
| 10:48 |
sfan5 |
woops kinda forgot |
| 11:03 |
|
Fixer joined #minetest-dev |
| 11:16 |
|
erlehmann joined #minetest-dev |
| 11:59 |
|
erlehmann joined #minetest-dev |
| 12:31 |
|
Edgy1 joined #minetest-dev |
| 12:33 |
|
Edgy1 joined #minetest-dev |
| 12:38 |
|
Edgy1 joined #minetest-dev |
| 13:58 |
|
gorbachev_pizza joined #minetest-dev |
| 14:48 |
|
Taoki joined #minetest-dev |
| 15:13 |
|
erlehmann joined #minetest-dev |
| 15:15 |
|
srifqi joined #minetest-dev |
| 15:55 |
|
srifqi joined #minetest-dev |
| 16:27 |
|
srifqi left #minetest-dev |
| 18:07 |
sfan5 |
reminder https://github.com/minetest/minetest_game/milestone/5 |
| 18:35 |
|
NetherEran joined #minetest-dev |
| 20:56 |
|
hecks joined #minetest-dev |
| 20:58 |
hecks |
I noticed some worrying errors related to PerlinNoise and noiseparams in a recent unrelated PR, who's been messing with it and why? |
| 20:59 |
sfan5 |
I'm afraid you'll need to be more specific |
| 21:01 |
hecks |
okay, the PerlinNoise constructor is complaining about noiseparam fields being strings |
| 21:01 |
hecks |
this is bad, you can't represent a map seed as a lua number |
| 21:03 |
hecks |
then it complains about not having a "position" and I don't know what it wants because there's no mention of it in lua_api.txt |
| 21:04 |
hecks |
the PR is #9969 and I assume this is something that was merged earlier; I had to disable any code dealing with PerlinNoise to test it |
| 21:04 |
ShadowBot |
https://github.com/minetest/minetest/issues/9969 -- Fix broken coloring of wielditems by dcbrwn |
| 21:05 |
hecks |
I'm just wondering if somebody has been "fixing" PerlinNoise when it wasn't really broken |
| 21:08 |
sfan5 |
we don't break features on purpose if that's your question |
| 21:08 |
sfan5 |
the string vs number thing should've been fixed by #9954 |
| 21:08 |
ShadowBot |
https://github.com/minetest/minetest/issues/9954 -- Lua API: Log incorrect parameter types as error by SmallJoker |
| 21:09 |
sfan5 |
however unless the perlin noise code does a manual string conversion "full" map seeds will not work anyway since neither lua_tonumber nor lua_tointeger work with 64-bit ints |
| 21:09 |
hecks |
I use FFI to do the 64 bit arithmetic and it's been working until now |
| 21:10 |
hecks |
however, a string was the only way to pass it intact |
| 21:16 |
hecks |
did you know that when you get the current world's noiseparams, it gives you strings |
| 21:16 |
sfan5 |
that's news to me |
| 21:16 |
PGimeno |
hecks: do you include the trailing LL in the string? |
| 21:16 |
sfan5 |
either way the PerlinNoise api only has 32-bit seeds: https://github.com/minetest/minetest/blob/master/src/noise.h#L107 |
| 21:18 |
hecks |
I don't include the LL, I just use ffi.C._strtoull, add the number, turn it back to a string; I guess I can use tonumber if the seeds are only 32 bit |
| 21:18 |
hecks |
but the arithmetic for adding seeds is definitely 64 bit |
| 21:18 |
hecks |
so naively adding them as numbers will rarely give you a working seed |
| 21:19 |
NetherEran |
If you use minetest.get_perlin() to get your noise instead of PerlinNoise(), the seed is interpreted as offset from the world seed, if that helps your case. |
| 21:19 |
hecks |
there must have been a reason why that function hasn't been working for me |
| 21:19 |
hecks |
I think it doesn't work in init |
| 21:20 |
hecks |
5.2 has been giving me problems with that too; the stable build never gives me noiseparams in init any more |
| 21:20 |
|
oil_boi joined #minetest-dev |
| 21:20 |
sfan5 |
and inside on_generated? |
| 21:20 |
hecks |
I need them in init |
| 21:20 |
hecks |
let me show yu |
| 21:22 |
hecks |
https://a.uguu.se/M8d0VaTBJdds_maptex.png |
| 21:22 |
hecks |
apparently assets generated in init will be enumerated by the server and sent to clients :) |
| 21:23 |
NetherEran |
oh that's pretty cool |
| 21:23 |
sfan5 |
filling the media cache runs after mod init yes |
| 21:23 |
hecks |
sfan made a PR that gives me a workaround to the init problem, but I may want to use map info for other things in init as well |
| 21:23 |
|
calcul0n_ joined #minetest-dev |
| 21:23 |
hecks |
such as placing points of interest (towns etc) on the map when the server is created |
| 21:24 |
hecks |
overall PerlinNoise has been a mess and now it's an even worse mess |
| 21:24 |
hecks |
due to what looks like very pedantic changes |
| 21:25 |
hecks |
I'll wait some days until all those PRs are merged before making it an issue |
| 21:26 |
sfan5 |
you can take master and merge 9969 into it to properly test it |
| 21:27 |
hecks |
I'll wait; an issue needs an actual version it applies to |
| 21:28 |
hecks |
overall I do not like the idea of type checking and errors or even warnings on coercion; lua doesn't work that way |
| 21:29 |
sfan5 |
again: lack of coercion is (was) a bug |
| 21:30 |
hecks |
but it's supposedly logging an error now? |
| 21:30 |
hecks |
spam not my stdout with such nonsense https://a.uguu.se/XgJk6GKHtRuN_stringnumber.png |
| 21:31 |
sfan5 |
it logs an erorr (not raises an error) when types do not fit after coercion |
| 21:31 |
sfan5 |
error* |
| 21:31 |
hecks |
ah, okay |
| 21:32 |
PGimeno |
anyway, if Lua is taking care of the type conversion then you're going to lose precision anyway |
| 21:32 |
hecks |
in that case it might actually be worth it to crash... but I don't wanna be pedantic :] |
| 21:33 |
oil_boi |
sfan5, I originally tried to move it to 569 but *pos_f always returns 0,0,0 |
| 21:34 |
oil_boi |
I will try / BS instead of /10 now |
| 21:35 |
sfan5 |
hm okay, then leave it where it is |
| 21:40 |
oil_boi |
*/ BS has even more precision :D Editing commit! |
| 21:50 |
oil_boi |
Why is "fabsf(speedJ.Y) <= 0.5f * BS" shorter sfan5? |
| 21:52 |
sfan5 |
it replaces speedJ.Y >= -0.5f * BS && speedJ.Y <= 0.5f * BS |
| 22:00 |
oil_boi |
Ok that's fixed now |
| 22:00 |
oil_boi |
You know what would be really cool? If we could walk down stairs and slabs :D |
| 22:01 |
sfan5 |
does minecraft have that |
| 22:03 |
oil_boi |
I dunno, let me check |
| 22:06 |
oil_boi |
No it doesn't actually |
| 22:07 |
oil_boi |
Mineshafts that go downwards using stairs would greatly benefit from them, it would certainly be a nice polish to add |
| 22:07 |
|
appguru joined #minetest-dev |
| 22:08 |
oil_boi |
I'll give the stair checking algorithm a check to see if I can invert it |
| 22:10 |
sfan5 |
also regarding m_can_jump: that variable also provides e.g. the "disable_jump" group (?) so if it's broken it should be fixed instead of left out |
| 22:11 |
oil_boi |
Oh okay, well, let me see something with it |
| 22:23 |
|
Wuzzy joined #minetest-dev |
| 22:30 |
oil_boi |
Ok sfan5 I added in a new Boolean called "m_can_bouncy_jump" So people can easier read what is going on with the bouncy jump code |
| 22:30 |
oil_boi |
Tested and working |
| 22:33 |
oil_boi |
I will now test the "(speed_f->Y <= 0.5f * BS)" stair fix |
| 22:36 |
oil_boi |
But the reason why I had it at 0.0f was because I tested infiniminer, MC, Terasology and that's how the physics worked as well |
| 22:38 |
oil_boi |
Yeah I can still see a jolt up with 0.5f * BS and 2.0 * BS :< |
| 22:58 |
hecks |
sfan5, can we talk about your dynamic asset PR |
| 23:33 |
|
mizux joined #minetest-dev |
| 23:41 |
|
proller joined #minetest-dev |
| 23:46 |
|
T4im joined #minetest-dev |
