Time  Nick      Message
01:31 paramat   #8417 is now ready
01:31 ShadowBot https://github.com/minetest/minetest/issues/8417 -- Stabilise 'day night ratio' to fix object brightness flicker by paramat
01:38 p_gimeno  hm...
01:40 p_gimeno  The instability probably comes from the formula for linear interpolation. That formula is not monotonic and has problems. I suggest an alternate formula here: http://math.stackexchange.com/questions/907327/accurate-floating-point-linear-interpolation
01:45 p_gimeno  applied to this case: return f < 0.5 ? values[i-1][1] + f * td0 : values[i][1] - (1.0 - f) * td0;
01:47 p_gimeno  sorry, make that 1.0f so there's no double rounding
01:49 p_gimeno  and 0.5f
01:49 p_gimeno  so with these fixes: return f < 0.5f ? values[i-1][1] + f * td0 : values[i][1] - (1.0f - f) * td0;
01:52 p_gimeno  hm, I've swapped the operands I think
01:57 p_gimeno  ah no I haven't, but it's not td0, it's like this: return f < 0.5f ? values[i-1][1] + f * (values[i][1] - values[i-1][1]) : values[i][1] - (1.0f - f) * (values[i][1] - values[i-1][1]);
02:07 paramat   what are the problems? that seems more intensive
02:11 paramat   hm not monotonic
02:11 paramat   i guess that might cause a blip when using very slow time speeds
02:11 p_gimeno  yeah it might oscillate at some other values
02:12 paramat   not sure, because the function converts to an integer when returning, so any blip might not have an effect
02:13 p_gimeno  also, it seems too easy a mistake to change the table and not realize the values below
02:13 paramat   maybe i can consider this in the follow up PR, which will be more suitable for such an improvement
02:14 p_gimeno  sure
02:15 paramat   that imperfect linear interpolation method is used a lot in MT and is more readable
02:15 p_gimeno  well, lerp(A, B, t) is even more readable
02:15 paramat   yes good point about changing the table values
02:16 paramat   i'll correct that
02:16 paramat   the monotonic issue, i guess any blip will be tiny? as it's due to flat imprecision?
02:17 paramat   *float
02:17 p_gimeno  yes but it might happen in the frontier of an integer if you're unlucky
02:17 paramat   hm yeah
02:18 p_gimeno  it was indeed happening at the endpoints
02:18 paramat   yeah =)
07:11 ANAND     https://github.com/minetest/minetest.github.io/pull/165
07:13 ANAND     !title
07:13 ShadowBot customize.html: Fix outdated intro by ClobberXD · Pull Request #165 · minetest/minetest.github.io · GitHub
12:55 p_gimeno  what is the 'high_precision_fpu' setting for? it seems unused to me
13:08 sfan5     given to irrlicht on init IIRC
13:09 sfan5     no idea what it does with that
13:10 p_gimeno  ah, thanks
13:44 nerzhul   hello team, i opened a PR about an android crash, i think it happens in some conditions when leaving world and/or app
14:14 p_gimeno  I see an issue but not a PR
14:31 nerzhul   oops :) yes an issue :D
16:49 nerzhul   i just published a tiny maintenance PR for master
16:49 nerzhul   #8425
16:49 ShadowBot nerzhul: Error: Delimiter not found in "Connection timed out."
16:49 nerzhul   #8425
16:49 ShadowBot https://github.com/minetest/minetest/issues/8425 -- Cleanups around the mapgen init by nerzhul
20:02 clavi     when using a Postgres backend, does it matter what locale the database is in?
21:04 sofar     clavi: no, not for map data anyway
21:25 clavi     ok thanks
23:36 paramat   !seen ShadowNinja
23:36 ShadowBot paramat: I saw ShadowNinja in #minetest-dev 1 year, 8 weeks, 4 days, 21 hours, 43 minutes, and 3 seconds ago saying "(The reason the server-side code is so bad is mainly because of compatability)"
23:36 paramat   god grief
23:36 paramat   *good