Minetest logo

IRC log for #minetest-dev, 2020-09-24

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

All times shown according to UTC.

Time Nick Message
00:07 Warr1024 joined #minetest-dev
00:44 jonadab joined #minetest-dev
01:46 MTDiscord <j​osiah_wi> Why does the build system use CMake 2.6? Is there a particular reason that such an outdated version is in use?
01:48 Foz joined #minetest-dev
01:56 T4im joined #minetest-dev
02:12 T4im joined #minetest-dev
02:13 search_social joined #minetest-dev
02:40 T4im joined #minetest-dev
05:18 Seirdy joined #minetest-dev
06:37 olliy joined #minetest-dev
08:00 proller joined #minetest-dev
08:00 ShadowNinja joined #minetest-dev
08:40 calcul0n joined #minetest-dev
08:52 sfan5 why outdated? you can use any recent version of cmake
09:02 ssieb joined #minetest-dev
09:32 p_gimeno joined #minetest-dev
10:20 rubenwardy That's just the minimum version of CMake, you can use any version newer than that
10:30 Fixer joined #minetest-dev
10:36 behalebabo joined #minetest-dev
11:13 calcul0n_ joined #minetest-dev
11:31 proller joined #minetest-dev
11:56 T4im joined #minetest-dev
11:59 proller joined #minetest-dev
12:18 Calinou out of curiosity, does the current build system actually work with CMake 2.6? :P
12:18 Calinou we probably want to bump the version to something like 3.4 at least
12:19 Calinou Ubuntu 16.04 (EOL in less than 1 year) has CMake 3.5, Ubuntu 18.04 has CMake 3.10
12:32 pgimeno I'm using cmake 3.16
12:52 celeron55 i seem to have an even older one, but 3.something anyway
12:55 celeron55 (debian stable has 3.13)
12:57 MTDiscord <j​osiah_wi> If you claim you support CMake 2.6, then the syntax needs to support it, so you can't use new features...
12:59 MTDiscord <j​osiah_wi> CMake doesn't have built in support for C++11 until version 3.1, for example.
13:01 MTDiscord <j​osiah_wi> I tried changing the minimum version to 3.16 and changing the way client libs were linked, (generator expressions instead of if clauses). The build system broke.
13:12 Taoki joined #minetest-dev
13:22 sfan5 we also set -std=c++11 manually in addition to set(CMAKE_CXX_STANDARD 11)
13:22 T4im joined #minetest-dev
13:22 sfan5 so that wouldn't be a reason 2.6 can't work, but I have no idea what the actual minimum cmake version is
13:24 MTDiscord <j​osiah_wi> Right. The syntax must work, because CMake honors that for older versions. e.g. I would get a syntax error if I tried to use a generator expression without changing the minimum version to at least 2.8
13:59 T4im joined #minetest-dev
14:17 Seirdy joined #minetest-dev
14:22 rubenwardy Newer cmake is nicer - things are based more on targets rather than setting compile flags
14:22 rubenwardy for example, you no longer use include_directory much - you instead depend on a target, which has include directories
14:22 rubenwardy plus  add_compile_options() is super nice
14:52 proller joined #minetest-dev
14:57 Seirdy joined #minetest-dev
15:04 Seirdy joined #minetest-dev
15:05 MTDiscord <j​osiah_wi> Indeed. I'm aiming for stuff like: cmake target_compile_features(minetest PRIVATE cpp_std_11) target_link_libraries(minetest     PRIVATE         $<$<BOOL:${USE_CURL}>:${CURL_LIBRARY}>
15:05 MTDiscord <j​osiah_wi> That might not have worked well with the relay. I'm not sure.
15:07 MTDiscord <j​osiah_wi> Ideally using a variable for library headers and stuff isn't good either. The Find script should provide an interface target with the correct includes set on it.
15:09 Seirdy joined #minetest-dev
15:19 rubenwardy for the first,  set(CMAKE_CXX_STANDARD 11) is better. Minetest isn't a library, so target compile features isn't needed
15:19 rubenwardy well, it's also not a problem for other things to be compiled as 11
15:26 MTDiscord <j​osiah_wi> I don't know whether setting the standard globally is better; it certainly isn't worse. Setting it on the target is more consistent though; if you do it that way for a library you might as well do it that way for an executable. I don't think you lose anything by doing it that way, but correct me if I'm wrong.
15:40 appguru joined #minetest-dev
16:14 jas_ joined #minetest-dev
16:17 Seirdy joined #minetest-dev
16:18 proller joined #minetest-dev
16:18 Seirdy joined #minetest-dev
16:19 Seirdy joined #minetest-dev
16:42 T4im joined #minetest-dev
16:59 pmp-p joined #minetest-dev
17:35 homthack joined #minetest-dev
17:51 homthack joined #minetest-dev
18:15 Lunatrius joined #minetest-dev
18:45 T4im joined #minetest-dev
19:22 lhofhansl joined #minetest-dev
19:23 lhofhansl Does anybody have some time to take a quick look at #10426 ?
19:23 ShadowBot https://github.com/minetest/minetest/issues/10426 -- Handle world mods and world game mods in the config dialog. by lhofhansl
19:28 Darcidride joined #minetest-dev
19:51 proller joined #minetest-dev
19:51 T4im joined #minetest-dev
19:58 fluxflux joined #minetest-dev
21:08 pgimeno nice one
21:10 pgimeno I'd have to review that, I don't see how it can be implemented without something like #6898
21:10 ShadowBot https://github.com/minetest/minetest/issues/6898 -- Allow distinguishing mods by modpack
21:11 sfan5 I had no idea world/<worldid>/game existed
21:12 pgimeno I think he meants worlds/ not world/
21:12 pgimeno hm, come to think about it, I don't think I took that into account in 6898
21:14 sfan5 I still didn't know that you could put a game into the world folder
21:15 MTDiscord <G​reenXenith> Huh https://github.com/minetest/minetest/blob/5.3.0/doc/lua_api.txt#L119-L128
21:16 MTDiscord <G​reenXenith> I guess it makes sense
21:17 MTDiscord <L​one_Wolf> CDB really needs to support downloadable worlds lol
21:23 lhofhansl joined #minetest-dev
22:00 T4im joined #minetest-dev
22:03 T4im joined #minetest-dev
22:11 T4im joined #minetest-dev
22:12 Icedream joined #minetest-dev
22:13 T4im joined #minetest-dev
23:26 homthack joined #minetest-dev
23:48 homthack joined #minetest-dev

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