Time  Nick       Message
13:17 rubenwardy https://quey.org/@moongaia/103947686435316073
13:17 rubenwardy sounds like the server freezing up?
13:23 sfan5      hard to guess without further debugging
13:23 sfan5      and further debugging is a waste until it's known to reproduce on 5.2
14:14 nerzhul    sfan5 nice startup for the emerge part, i was thinking around it too, i commented you some improvements because i'm not sure we need to use so many heap and stack can suffice in many cases
14:15 sfan5      thanks for reviewing, will take a look
14:15 kilbith    sfan5 is on fire guys
14:15 nephele    get the extinquisher then!
14:24 sfan5      hm build failed ...
14:24 sfan5      ....and it's clang-format complaing about the formatting of the unittesting code
14:29 rubenwardy I'm very to happy sfan5's fire, I just hope he doesn't burn out to quickly
14:30 sfan5      oh don't worry I will run out of things to do soon enough
14:44 nerzhul    but nice work, i may continue some things after the merge, i'm working on some very old code refactor with no function changes because we tend to develop things based on issues but not get time to refactor the existing one
15:32 sfan5      reminder: #9586 is small and still needs a review
15:32 ShadowBot  https://github.com/minetest/minetest/issues/9586 -- Implement DPI scaling for Windows by sfan5
15:35 nerzhul    commented
16:04 nerzhul    i just published https://github.com/minetest/minetest/pull/9629/files refacto PR. It should also improve a little bit the active object messages performance but removing some copy due the old non AO ownershipped code
16:21 nerzhul    the coredev team is the same than before ? nobody more to help us ? :) time to take some minutes to do refactor on the code while working around new user needs or bug to fix
16:22 sfan5      https://github.com/orgs/minetest/teams/engine/members
16:28 nerzhul    yep just look at is, very stable :D
16:29 nerzhul    sfan5: answered your comments
16:44 nerzhul    travis is green fine, i was not sure due to emplace on std::queue, but it's fine, great
16:44 sfan5      I'd expect emplace to work pretty much everywhere if push exists
16:44 nerzhul    yep :D
16:46 nerzhul    i think at a point we can improve the whole serialization but it's a huge refactor for later, because we did stream => string => stream/sth else
16:46 nerzhul    let me a second i will add a second commit to protect m_messages_out which is not only called in case of pop
16:47 nerzhul    now*
16:51 sfan5      the static lifetime wasn't obvious at all unless you also look at the .h
16:52 nerzhul    you prefer that i put both const and non static ? i agree it is easier to understand, i can amend with unifying the way to create messages, ok ?
16:52 sfan5      static is okay like this
16:53 nerzhul    okay then i don't touch
16:54 nerzhul    i just push in few minutes the m_message_out attribute protection
16:57 nerzhul    okay the second commit to protect the SAO queue is pushed
16:57 nerzhul    feel free to review & approve :)
16:58 nerzhul    tell me when you want another review on your emerge thread optimization
17:13 sfan5      probably after paramat has had a look at it
17:14 nerzhul    on your PR you mean ? yeah i think so, the optimization is fine, as we don't change the objects at runtime
17:16 sfan5      yes
17:22 nerzhul    just re-tested my PR with the new addition it's fine, we should have a tiny eprformance improvement when many SAO have messages for us
17:22 Wuzzy      btw what are the plans for 5.3.0? I heard you wanted to release it shortly after 5.2.0
17:23 sfan5      https://github.com/minetest/minetest/milestone/16 the milestone description sheds some light on that
17:25 Wuzzy      thx
17:25 Wuzzy      ah, right, the list makes total sense
17:26 nerzhul    shortly = 6 months like always or ? :D
17:27 nerzhul    very very trivial MR (move a header only file to the right folder) https://github.com/minetest/minetest/pull/9630
17:28 sfan5      lgtm
17:29 nerzhul    ty
17:29 Wuzzy      This PR is 1 month old and has no labels: https://github.com/minetest/minetest/pull/9453 ... was it overlooked?
17:31 sfan5      even PRs with labels get overlooked ;)
17:31 Wuzzy      yeah, with >150 open prs, this happens :(((
17:32 Wuzzy      i think the majority of PRs have no chance in snowball's hell to be merged, especially the ancient ones. i guess a lot of them need to be killed off (sadly)
17:33 Wuzzy      the fact that there are so extremely many open prs makes development not exactly easy, i suppose ...
17:33 Wuzzy      are there any criteria on when to kill a PR?
17:34 Wuzzy      I have the impression that many PRs are just left open (and rotting) forever.  like in purgatory. they neither go to upstream heaven or to rejection hell
17:48 nerzhul    yeah what happened with PR number :D on my afk we tried to be around 100 and now we have > 160 :D
17:49 nerzhul    merging #9629
17:49 ShadowBot  https://github.com/minetest/minetest/issues/9629 -- Drop genericobject.{cpp,h} by nerzhul
17:49 nerzhul    thanks sfan5 for the review
18:10 nerzhul    sfan5 i merge #9630, build is green on all platforms
18:10 ShadowBot  https://github.com/minetest/minetest/issues/9630 -- Move clientsimpleobject.h to client folder by nerzhul
18:10 nerzhul    as you approved it here
18:16 sfan5      sure
18:26 nerzhul    merging #9579
18:26 ShadowBot  https://github.com/minetest/minetest/issues/9579 -- Print error message if an invalid mapgen alias was detected by Wuzzy2
19:03 nerzhul    rubenwardy, sfan5, what do you think about spliting the SAO by type in files ? the current content_sao.cpp file is just crazy in terms of code lines. I can take time to do the split this weekend
19:03 rubenwardy sgtm
19:04 rubenwardy there was a PR to do that before, no?
19:04 sfan5      sounds useful yes
19:05 nerzhul    i think i did a try at a moment but it didn't succeed cannot remember why
19:05 nerzhul    maybe due to our strange SAO factory :D