Minetest logo

IRC log for #minetest-dev, 2020-12-30

| Channels | #minetest-dev index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
03:05 olliy joined #minetest-dev
05:00 MTDiscord joined #minetest-dev
05:34 YuGiOhJCJ joined #minetest-dev
08:00 ShadowNinja joined #minetest-dev
08:39 Thomas-S joined #minetest-dev
08:39 Thomas-S joined #minetest-dev
09:09 lisac joined #minetest-dev
09:11 calcul0n_ joined #minetest-dev
09:17 Edgy1 joined #minetest-dev
09:20 Calinou joined #minetest-dev
09:22 calcul0n joined #minetest-dev
09:49 proller joined #minetest-dev
10:26 calcul0n_ joined #minetest-dev
12:24 Fixer joined #minetest-dev
12:39 T4im joined #minetest-dev
12:43 calcul0n__ joined #minetest-dev
13:29 T4im joined #minetest-dev
15:50 fluxflux joined #minetest-dev
16:06 hecks joined #minetest-dev
18:33 lhofhansl joined #minetest-dev
18:34 lhofhansl Hello all. I would appreciate a closer review of #10709 if someone has the time.
18:34 ShadowBot https://github.com/minetest/minetest/issues/10709 -- Allow map save to compress blocks asynchronously. by lhofhansl
18:36 homthack joined #minetest-dev
18:37 Andrey01 joined #minetest-dev
18:41 sfan5 soon(tm)
18:42 Krock why do you need swapSerializedBlocks? would it be possible to &&-move the ownership after swapping to a new list? this approach is kinda obscure
18:43 Krock I understand that this reduces the computation efforts, but I wonder whether there's not a better solution than feeding a dummy object
18:43 Krock s/dummy/new/
18:44 Krock also savemap_timer could be replaced by IntervalLimiter
18:47 celeron55 eh, lol
18:47 celeron55 i went to read IntervalLimiter's implementation
18:48 celeron55 it's so simple, yet the only comment got the return value completely mixed
18:49 lhofhansl Krock: I had a different version first where timerUpdate and save are passed a list that would be filled with the results. This is way it is more contained inside Map.
18:49 celeron55 and it's written by me
18:50 lhofhansl Krock: There's locking or making a copy and then work on the copy. I do not see any other approach. Locking would likely need to be too fine grained.
18:51 lhofhansl I tried a copy constructor for MapBlock but that spiraled out of control quickly - perhaps I can try again.
18:52 lhofhansl and the savemap timer I kept from the original. Happy to change it obviously.
18:53 lhofhansl let's discuss on the PR.
19:00 proller joined #minetest-dev
19:05 Krock I wish NRVO were guaranteed
19:16 hecks joined #minetest-dev
19:40 lisac joined #minetest-dev
19:59 lhofhansl Krock: Me too. Hence the weird swap approach. I'll address your comment on the PR today.
22:09 appguru joined #minetest-dev
23:14 sofar rubenwardy: now i see where that media server question came from,  lol

| Channels | #minetest-dev index | Today | | Google Search | Plaintext