Time  Nick      Message
06:53 est31     pushing in 10 minutes : https://github.com/est31/minetest/commit/8df89db30e80eff2ebbf30f0331b8ab4906a410e
07:22 paramat   game#1238 updated
07:23 ShadowBot https://github.com/minetest/minetest_game/issues/1238 -- Default: Optimise and simplify leafdecay ABM, remove cache by paramat
07:41 est31     announcing flatpak build script for minetest https://github.com/est31/minetest-flatpak
08:09 est31     about the flatpak build script: it is best to be in a combination with an official repo where we can push the app to.
08:09 est31     possibly even with nightly builds
08:10 est31     for this however, a build server is required. Anyone want to donate the service?
08:10 est31     I can't tell what exact system requirements there are, but it needs to be linux, and recent enough to support flatpak
08:11 est31     hosting can possibly be done over github
08:12 nrzkt     est31 why a build server if there is travis ?
08:14 est31     hrmm that's an idea
08:32 nrzkt     est31, i miss something, what do you want to do ? i only have the question :D
08:33 est31     nrzkt, I want to create an official minetest flatpak repository
08:33 nrzkt     i see
08:34 nrzkt     flatpak is fedora, right ?
08:34 est31     it is made to run on any linux desktop
08:34 est31     but yes, originally fedora made it
08:34 est31     it will be available in ubuntu 16.10 though
08:35 est31     I have it through a ppa :)
08:35 est31     so I've done the build script https://github.com/est31/minetest-flatpak
08:35 est31     now I only need to set up the infrastructure
08:35 nrzkt     but ubuntu have their own thing no ?
08:35 est31     yes, ubuntu has snappy
08:35 nrzkt     the repository is an application ?
08:36 est31     as far as I understand it, each repository can have multiple applications
08:36 nrzkt     like docker registry then ?
08:36 est31     I think so
08:36 est31     a repository is basically an url plus a signing key
08:37 est31     about snappy vs flatpak, I believe that flatpak has the better chances to win the fight
08:38 est31     but both are better than appimage in my eyes, and all we can do is to bet on one horse and hope the best :)
08:47 est31     hm, my repo is actually > 100 mb large
08:47 est31     thats big, for only one app
08:48 est31     hrmm when I clean it up it only needs 25 mb
08:54 nrzkt     est31, => https://github.com/minetest/minetest/pull/4430
08:55 nrzkt     est31, not large because you have all MT assets and deps
08:55 est31     does that need an entirely new file?
08:55 nrzkt     i think, to prevent adding map header everywhere
08:56 est31     hrmm that makes sense
08:56 nrzkt     it replace the <map> header for unodered_map only cpp files
08:57 nrzkt     this pr will permit to improve performance by changing many std::"ordered_"map to unordered, there are very few places where we need order
09:00 nrzkt     sfan5, sofar, nore can you look at #4430 ?
09:00 ShadowBot https://github.com/minetest/minetest/issues/4430 -- Add a header to permit using std::unordered_map on modern c++11 and upper compilers by nerzhul
09:02 nrzkt     est31 i think we should have a forced c++11 compiler using gcc 6.0 or 5.1 on travis, because atm we didn't know which c++ version is used on travis
09:03 nrzkt     like here: https://github.com/spacel/game/blob/master/.travis.yml
09:08 Krock     a whole new file for this definition?
09:09 nrzkt     Krock, i justified the usage before
09:10 Krock     oh my bad. Haven't read the discussion above
09:10 nrzkt     it replace <map> header in many files, else the solution is to include <map> and <unordered_map> everywhere and it's not proper :)
09:12 Krock     I see. a-ok then
09:25 sfan5     nrzkt: i suggest calling the header c11containers.h or similar
09:25 sfan5     it could also be used for unordered_set for example
09:25 est31     then every time you want a map you will include unordered_set as well?
09:26 nrzkt     sfan5, the est31 issue was the issue i want to prevent here
09:26 est31     there is already a file it would sort of fit in, src/util/container.h
09:26 nrzkt     est
09:27 sfan5     micro-optimizations (especially at compile time) of including a single file more are useless
09:27 nrzkt     est31 container.h is not included everywhere
09:27 sfan5     container.h sounds fitting though
09:27 sfan5     why not move it there?
09:27 nrzkt     because everything doesn't include container.h
09:28 nrzkt     you will include list, vector, map, set, queue, irrlicht, exceptions, mutex and semaphores everywhere
09:29 nrzkt     are your sure it's what you want ? est31 , sfan5 ?
09:29 est31     nrzkt, sfan5 both are okay for me
09:30 sfan5     >irrlicht everywhere
09:30 sfan5     hmm no
09:31 sfan5     nrzkt: the c11container.h approach sounds best to me
09:31 sfan5     cpp11*
09:31 nrzkt     sfan5, okay then i rename the file to cpp11 and add same thing for std::unordered_set
09:32 sfan5     % grep -RF "std::set<" . | wc -l
09:32 sfan5     230
09:32 sfan5     that's a lot of sets
09:32 nrzkt     yes but if i remember there is many sets which needs to be ordered :p
09:33 sfan5     someone will have to go through the sets later
09:34 nrzkt     maps are easier to change, 80% of them needs to be ordered_map, 15% are v3 or v2 then it's not possible to use unordered_set except if we override the class to have the hashable compliance
09:35 sfan5     hashing a vector shouldn't be hard
09:45 nrzkt     est31, sfan5 #4430 updated (description too)
09:45 ShadowBot https://github.com/minetest/minetest/issues/4430 -- Permit usage of std::unordered_map & std::unorderered_set on c++11 compilers by nerzhul
09:45 est31     can it be named cpp11_container.h ?
09:46 est31     c11 is a c standard :=)
09:46 nrzkt     yes it's possible rmrlmlr :p
09:46 est31     plus underscore, Tesseract likes those
09:48 nrzkt     done
09:49 est31     aaand I think we can merge it
09:49 est31     nrzkt, have you tested whether it builds?
09:50 Krock     travis is in progress
09:50 est31     lets wait for travis
09:51 nrzkt     it builds without problems yes :p i have tested it in a world to ensure this doesn't have side effects :p
09:51 nrzkt     "new world"*
09:51 nrzkt     i build everytime before i push (i have a 8 core CPU it's very fast, < 40sec for all MT)
09:52 Krock     I'm jealous
09:52 nrzkt     :p
09:53 nrzkt     don't hesitate to replace some not sensible maps and sets when you get some time, there are many places to optimize using this :p
09:53 nrzkt     when i get motivation i will finish my PR for player attributes...
09:58 nrzkt     est31, for the flatpack issue i'm not sure they win, i see many many interesting projects on twitter and they communicate about snaps, docker but not flatpack :s (owncloud, gitlab...)
09:58 est31     flatpak is no replacement for docker
09:58 est31     docker is meant for servers, flatpak for desktops
10:00 est31     and about flatpak vs snappy, from what I've heard snappy is in the early stages of development with lots of stuff not working yet
10:01 est31     but flatpak is quite mature already
10:03 est31     nrzkt, what about we offer flatpak for now, and if someone has time to set it up, snappy as well :)
10:04 nrzkt     i know this is not a replacement, i said libre product's enteprises seems to use docker and snappy
10:05 nrzkt     est31, no problem for me :) if i get some time a day i will work on a strong docker image for MT
10:05 est31     https://www.libreoffice.org/download/snap/
10:05 est31     https://www.libreoffice.org/download/flatpak/
10:05 est31     they offer both :)
10:05 nrzkt     good :)
10:06 nrzkt     build pass :)
10:13 Krock     If only my pulls would get merged so fast :3
10:15 est31     yeah there is a huge overhead of the PR queue
10:15 est31     someone has to go through it and review them all
10:16 nrzkt     est31, #4400 is okay for me too
10:16 ShadowBot https://github.com/minetest/minetest/issues/4400 -- Add zoom with new "zoom" privilege by bendeutsch
10:17 nrzkt     #4408 too
10:17 ShadowBot https://github.com/minetest/minetest/issues/4408 -- Lua->C getintfield() use lua_tointeger by tomasbrod
10:20 est31     all merged
10:21 nrzkt     ty
12:21 nrzkt     will merge #4313 soon
12:21 ShadowBot https://github.com/minetest/minetest/issues/4313 -- Supplement build steps for Windows by DLaboratory
12:52 nrzkt     sfan5, nore can you review #4364 it's important
12:52 ShadowBot https://github.com/minetest/minetest/issues/4364 -- couple of memory leaks fixes. by devnexen
12:52 nrzkt     i added my approval
13:00 sfan5     nrzkt: added my approval too
14:55 nrzkt     sfan5, ty
16:05 Fixer     observation: there are lots of people "joining and time outing" all the time on numerous servers (incl light ones), i have no idea why (maybe players on 128mb ram phones?)
16:29 T4im      Fixer: mobile connections in combination with huge assets to download?
16:29 T4im      people might just abort
16:29 Fixer     T4im: huge assets? it was literally default mtg
16:29 Fixer     how small it can be?
18:33 Fixer     https://github.com/minetest/minetest/pull/3810 needs action
18:40 Krock     like any other pull
19:11 sofar     est31: do you actually have a minetest flatpak somewhere? I'd like to test it
20:13 APNG      hi, so I wanna promote APNG
22:22 nrzkt     merging #4431 as it's trivial an has two approvements
22:22 ShadowBot https://github.com/minetest/minetest/issues/4431 -- Add infotext containing entity type (e.g. mobs:cow) to unknown entities by Rogier-5
23:54 est31     sofar, I have built the flatpak file now, any idea where I could upload it?
23:55 est31     its semi official, I dont want to attach it to the release just yet
23:59 sofar     add it to a gituhb tag ?