| Time |
Nick |
Message |
| 00:21 |
|
Noisytoot joined #minetest-dev |
| 01:40 |
|
ShadowBot joined #minetest-dev |
| 02:01 |
|
ShadowBot joined #minetest-dev |
| 04:00 |
|
MTDiscord joined #minetest-dev |
| 08:14 |
|
YuGiOhJCJ joined #minetest-dev |
| 12:03 |
|
flux__ joined #minetest-dev |
| 13:38 |
|
Lupercus joined #minetest-dev |
| 13:40 |
|
Lupercus joined #minetest-dev |
| 15:10 |
|
Fenrir241 joined #minetest-dev |
| 15:13 |
|
Fenrir241 left #minetest-dev |
| 16:07 |
|
Lupercus joined #minetest-dev |
| 16:30 |
sfan5 |
merging #14563, #14579, #14556, #14576 in 25m |
| 16:30 |
ShadowBot |
https://github.com/minetest/minetest/issues/14563 -- [no sq] Unnamed collection of fixes by sfan5 |
| 16:30 |
ShadowBot |
https://github.com/minetest/minetest/issues/14579 -- Remove an unnecessary copy of a big data structure. by cosin15 |
| 16:30 |
sfan5 |
15m* |
| 16:30 |
ShadowBot |
https://github.com/minetest/minetest/issues/14556 -- Formalize OS support policy by sfan5 |
| 16:30 |
ShadowBot |
https://github.com/minetest/minetest/issues/14576 -- Fix MSVC warning C4172 in ModifySafeMap::get by JosiahWI |
| 17:42 |
|
Lupercus joined #minetest-dev |
| 18:44 |
|
Noisytoot joined #minetest-dev |
| 18:46 |
|
Fenrir24 joined #minetest-dev |
| 19:16 |
sfan5 |
I wonder if it would be useful if an entity could request to only be stepped every 0.5s instead every server step |
| 19:16 |
sfan5 |
could save a lot of calling into lua |
| 19:26 |
rubenwardy |
Light sleep mode |
| 19:27 |
sfan5 |
0.5s is an example of course, the entity would decide the value |
| 19:28 |
Mantar |
could be useful, especially if you could specify the delay. our spiders in Exile can go into a hibernation mode where they go quiet and motionless and wait for prey to move into range, to conserve energy |
| 19:29 |
Mantar |
would seem sensible to lower their step rate too |
| 19:32 |
MTDiscord |
<landarvargan> A lot of entities I've made in my time modding don't even use on_step |
| 19:33 |
MTDiscord |
<landarvargan> A mode where an entity only steps on_collide would be interesting |
| 20:33 |
|
Noisytoot joined #minetest-dev |
| 21:10 |
|
Noisytoot joined #minetest-dev |
| 21:35 |
MTDiscord |
<luatic> have we determined that the overhead of calling into lua just to have the entity check a timer is a significant performance problem for entities? |
| 21:47 |
|
turtleman joined #minetest-dev |
| 21:54 |
sfan5 |
no, but I suspect it |
| 22:18 |
|
Fenrir24 joined #minetest-dev |
| 22:19 |
sfan5 |
instead of that however it'd be much more impactful if some big server owner shared perf data of their prod setup |
| 22:19 |
sfan5 |
doesn't rubenwardy run the ctf one? |
| 22:27 |
MTDiscord |
<luatic> @landarvargan runs it now |
| 22:33 |
|
panwolfram joined #minetest-dev |
| 22:53 |
MTDiscord |
<landarvargan> Yeah. I'm willing to help find perf issues and test fixes for them. Would this perf data be profiler dumps? |
| 22:55 |
MTDiscord |
<luatic> in this case yes. that should also help us identify server performance issues more generally. |
| 23:10 |
|
proller joined #minetest-dev |
| 23:11 |
proller |
maybe just use c++ instead lua where performance required? |
| 23:15 |
MTDiscord |
<luatic> proller: that is what sfan was suggesting (avoiding the timer in lua), but first we need to determine where performance is required, and for that we need perf data |
| 23:31 |
|
Lupercus joined #minetest-dev |
| 23:40 |
|
Fenrir241 joined #minetest-dev |
