Time  Nick       Message
10:54 ANAND      Weird... std::cout in Client::sendPlayerPos (called every step) prints nothing to the console.
10:56 ANAND      I'll try using warningstream
10:57 sfan5      i just use actionstream when i need to debug anything
11:24 pmp-p      is WM app focus handled internally by irrlicht or in minetest code ? even looking at bits of android code i don't see anything about pause/resume
11:28 sfan5      most likely irrlicht
11:43 ANAND      sfan5: I use cout, so that the messages can be formatted the way I like - I also add a couple of \n's to make them easier to spot. :)
11:43 ANAND      But for whatever weird reason, warningstream works...
11:48 pmp-p      yeah probably irrlicht and some probleme with the version i use (1.8.3) switching task takes 4 seconds even if artificially lowering app load with sleep_ms calls
15:26 rubenwardy .https://www.reddit.com/r/minetest/comments/ctyrdc
17:20 Krock      32 new lines for the occlusion culling code. let's see how that works
17:25 rubenwardy Proposal for next meeting: what do we want to have in 5.1.0?
17:25 rubenwardy !dev Meetings
17:25 ShadowBot  Meetings - Minetest Developer Wiki -- http://dev.minetest.net/Meetings
17:25 sfan5      what do you mean by "have"?
17:25 sfan5      in the sense of, do we want any specific features added before 5.1 is released?
17:25 sfan5      another point to dicuss would be when to release 5.1
17:26 rubenwardy yeah. I think that it would be good to focus a lot on formspecs for 5.1, as we have been doing
17:26 rubenwardy I propose November/December for 5.1.0
17:26 rubenwardy as, in true MT fashion, november won't happen but makes december more likely :)
17:27 sfan5      we're allowed to release more than once a year btw
17:27 rubenwardy that's twice a year
17:27 rubenwardy oh wait, we released february
17:28 rubenwardy our typical (before 5) schedule was May-July and December
17:28 sfan5      what i'm trying to say is that i'd prefer releasing more like every 6 mo than more like every 12
17:29 rubenwardy yeah, that would be my preference
17:30 sfan5      then why not move the schedule so it fits? Aug-Oct and March
17:30 rubenwardy sounds good to me
17:30 sfan5      (well it doesn't need to be exactly that schedule)
17:30 rubenwardy if we had the resources to do so, my preference would be a minor release every 6 months and a patch every 1-2 months
17:30 rubenwardy but: limited time, and not always bug fixes that make it worth it
17:30 rubenwardy Minetest 5.0 was March 1st
17:30 sfan5      master is supposed to be mostly stable so I don't see value in patch releases
17:31 rubenwardy heh
17:31 rubenwardy the aim would be to increase the stability without adding new features to decrease it
17:31 rubenwardy but I see your point
17:31 rubenwardy also, we're a game
17:31 rubenwardy not an OS
17:33 sfan5      rubenwardy: the old schedule was more frequent at times https://a.uguu.se/Iy0qWZd51pPC_DTriR5Y.png
17:33 rubenwardy *4th
17:33 sfan5      in the end it's down to how much time is available
17:34 rubenwardy yeah, it varied in interval but was twice a year on average
17:34 rubenwardy that's true
17:34 rubenwardy I have a lot of time before September 16th, as that's when I start my new job
17:36 rubenwardy I've mostly been trying to work out how to graphics   https://rwdy.uk/eIYDG.png
17:41 sfan5      Krock: did you push your occlusion changes yet?
17:42 Krock      the newest not-worse-than-in-master code is up, but the experimental node searching stuff isn't
17:42 Krock      I can sometimes watch through long corridors now, but it sometimes skips one mapblock
17:56 Krock      regarding experimental: https://pastebin.com/raw/0Qf18tGf  this is what I had in mind
18:03 sfan5      take the 5th mapblock in the shootline to the target, average the position of "all" air nodes, extrapolate that to the target mapblock, then do the occlusion check
18:03 sfan5      do i have that correct?
18:07 Krock      not mapblock, but just nodes
18:08 Krock      but the rest is right. using the changed direction vector for the new occlusion target mapblock
18:12 sfan5      that seems quite complicated, especially with the 27 node lookups
18:13 Krock      used 27 to get more precise average values, but 7 might work as well
18:17 Krock      after all it does not have a noticeable impact