Minetest logo

IRC log for #minetest-dev, 2015-09-05

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

All times shown according to UTC.

Time Nick Message
00:05 VanessaE paramat:  can you update your saveschems repo to reflect your latest tree changes, please?
00:24 leat joined #minetest-dev
00:30 kaeza joined #minetest-dev
01:19 rickmcfarley joined #minetest-dev
02:57 Samson1 joined #minetest-dev
03:21 Samson1 joined #minetest-dev
03:27 FR^3 joined #minetest-dev
03:32 DFeniks joined #minetest-dev
03:33 FR^3 joined #minetest-dev
03:44 FR^3 joined #minetest-dev
04:49 paramat joined #minetest-dev
04:50 paramat i noticed an error in the jungletree from sapling schem, fixing in a moment
04:51 VanessaE [09-04 20:05] <VanessaE> paramat:  can you update your saveschems repo to reflect your latest tree changes, please?
04:51 paramat is done =)
04:51 VanessaE :D
05:03 T4im joined #minetest-dev
05:06 paramat game#664
05:06 ShadowBot https://github.com/minetes​t/minetest_game/issues/664 -- Default/schematics: Fix root error in jungletree from sapling by paramat
05:06 paramat now merging
05:11 paramat complete
05:12 VanessaE don't forget your saveschems repo again :)
05:13 paramat that already has the fix in it
05:13 VanessaE ok
05:20 Samson1 joined #minetest-dev
05:21 diemartin joined #minetest-dev
05:47 paramat left #minetest-dev
05:58 younishd_ joined #minetest-dev
06:05 leat joined #minetest-dev
06:32 Calinou joined #minetest-dev
06:52 blaze joined #minetest-dev
07:02 Hunterz joined #minetest-dev
07:26 svenstaro so
07:26 svenstaro you guys all had a chance to test the new arhclinux packages
07:27 svenstaro does it work ok for you?
07:37 CraigyDavi joined #minetest-dev
07:45 sfan5 svenstaro: i can test it if you tell me where to get it
07:50 svenstaro sfan5: normal arch way
07:50 svenstaro pacman -S minetest
07:50 svenstaro minetest
07:50 svenstaro crash or not?
08:01 Yepoleb_ joined #minetest-dev
08:13 Samson1 joined #minetest-dev
08:15 sfan5 svenstaro: no crash
08:22 technics joined #minetest-dev
08:25 svenstaro sweet
08:29 Dartmouth joined #minetest-dev
09:14 younishd joined #minetest-dev
09:29 CraigyDavi joined #minetest-dev
09:34 CraigyDavi joined #minetest-dev
09:42 Megaf svenstaro: Hi there, hows things?
09:45 Megaf #3130
09:45 ShadowBot https://github.com/minetest/minetest/issues/3130 -- segfault on startup for irrlicht 1.8.2 on arch linux
09:45 Megaf svenstaro:  irrlicht still broken
09:55 sfan5 uh
09:55 sfan5 Megaf: works for me with 0.4.12-2 from the repos
09:56 sfan5 also the minetest i build from source works with the normal irrlicht
09:57 sfan5 % pacman -Q irrlicht
09:57 sfan5 irrlicht 1.8.1-2
09:58 svenstaro sfan5: so it works for you now in all configurations?
09:59 svenstaro as long as you use irrlicht from the repos
09:59 sfan5 yes
09:59 sfan5 both minetest from the repos and the one i compiled myself work
10:07 Puma_rc joined #minetest-dev
10:14 est31 joined #minetest-dev
10:15 est31 sfan5, do you use gcc or clang for compilation?
10:15 est31 I guess clang, no?
10:15 sfan5 yes clang
10:16 est31 can you do a re-clone and try with gcc
10:16 est31 because thats the theory
10:16 sfan5 sure
10:16 est31 svenstaro has just changed compilers for the irrlicht headers
10:16 est31 err
10:16 est31 irrlicht package*
10:17 est31 so its probable that if minetest still fails it might be a compiler problem as well
10:17 est31 and the problem lurking inside irrlicht headers
10:20 Megaf est31: other topic, I noticed that cpp files for leveldb and gettext and perhaps redis and some others are compile even tho they are disabled with cmake
10:20 Megaf s/compile/compiled
10:21 est31 well they are compiled yes
10:21 Megaf fixing CMakeLists.txt on src with some ifs might solve that
10:21 est31 but they consist of an #ifdef USE_REDIS at the start
10:21 est31 and an #ifdef at the end
10:21 est31 so it shouldnt matter too much
10:21 Megaf why are we wasting time and cpu compiling them if they are not going to be used?
10:22 est31 they aren't compiled
10:22 sfan5 est31: nope works
10:22 Megaf sfan5: define nope works
10:22 sfan5 works with gcc too
10:22 est31 which gcc version?
10:22 Megaf ok, so we have to go deeper
10:23 sfan5 5.2.0
10:23 est31 okay theory busted then
10:23 * Megaf blames Arch/Kernel once again
10:23 Megaf xP
10:24 Megaf sfan5: what kernel are you on?
10:24 sfan5 4.1.5-1-ARCH
10:24 est31 this means your arch system is fully up to date?
10:24 est31 dont know what version numbers mean on rolling release
10:24 Megaf sfan5: other guy with the problem was using the Zen kernel
10:25 sfan5 est31: version numbers mean exactly the same
10:25 sfan5 and no im not up to date
10:25 Megaf you are not using the Zen kernel and it works for you
10:25 sfan5 havent upgraded to linux 4.1.6-1 yet
10:25 svenstaro it's not
10:25 svenstaro 4.1.6 is new
10:25 Megaf $ uname -a
10:25 Megaf Linux ToshSam 4.1-6.dmz.1-liquorix-amd64 #1 ZEN SMP PREEMPT Debian 4.1-9 (2015-08-17) x86_64 GNU/Linux
10:25 svenstaro 4.2 in testing
10:26 svenstaro someone report a bug to irrlicht
10:26 est31 sfan5, was it broken for you before the 1.8.2-2 upgrade?
10:26 Megaf well, it works on 4.1.6, but I'm using GCC 4.9
10:26 svenstaro I don't like sourceforge
10:26 sfan5 it was never broken for me
10:26 Megaf svenstaro: nobody does
10:26 est31 but you had 1.8.2-1?
10:27 sfan5 [2014-07-17 20:54] [PACMAN] installed irrlicht (1.8.1-2)
10:27 Amaz joined #minetest-dev
10:28 est31 ah you still have 1.8.1?
10:28 sfan5 yes
10:28 est31 well, then you arent affected.
10:28 est31 the bug is with 1.8.2
10:28 sfan5 I'm updating right now
10:28 sfan5 i can try with gcc and clang again
10:31 est31 at least we know now that gcc 5.2 with 1.8.1 is perfectly fine
10:31 sfan5 clang 3.6.2 too
10:31 svenstaro mh
10:32 svenstaro I'm gonna compile minetest with clang for fun and put it into arch
10:32 sfan5 my minetest compiled with 1.8.1 headers still works
10:32 svenstaro maybe it's the headers after all
10:35 svenstaro actually no I'm not as it's not my package
10:37 est31 I wonder, if anybody can reproduce the crash with irrlicht 1.8.2-2, and the binary 0.4.12 minetest package
10:38 sfan5 i can't
10:38 sfan5 irrlicht 1.8.2-2 and minetest 0.4.12-2 works
10:39 sfan5 compiled from source works though
10:39 sfan5 s/works/crashes/
10:39 sfan5 that means it's probably the 1.8.2 headers
10:39 RealBadAngel 1.8.2-2 was supposed to be fine
10:39 Player_2 joined #minetest-dev
10:39 est31 well, the 1.8.2-2 binary is fine
10:40 est31 the 1.8.2-2 headers arent, as it seems
10:40 est31 sfan5, which compiler did you try?
10:40 sfan5 gcc
10:41 sfan5 same backtrace as in the issue
10:41 est31 hrmmm, is getting another gcc version as easy as downloading the tarball, extracting doing ./configure && make and this is it?
10:42 sfan5 you forgot make install but yes
10:42 est31 or do I need toolchain, libstd etc as well?
10:42 sfan5 thats included in the gcc source
10:42 sfan5 but you probably want to use a build script that builds a pkg for your distro to make sure you break the gcc package
10:42 sfan5 +don't
10:43 est31 well, I'll let gcc set its "install" dir into /home/$somewhere
10:50 sfan5 huh strange
10:50 sfan5 compiling minetest with clang (w/ the irrlicht 1.8.2-2) does not lead to a crash
10:51 est31 thats what should happen, after the theory
10:51 est31 1.8.1-1 was completely broken, binary and headers so that even nrzkt while he uses clang couldnt start the binary
10:52 est31 but 1.8.2-2 seemed to fix it for him
10:52 sfan5 <est31> 1.8.1-1 was completely broken,
10:52 sfan5 exact opposite for me
10:52 sfan5 1.8.1-1 always worked fine
10:52 est31 ah sorry
10:52 est31 meant 1.8.2-1
10:52 sfan5 ah
10:54 est31 sorry, it seems I've read nrz wrongly, he hasnt said that
10:54 svenstaro uhm
10:54 svenstaro so basically clang is k
10:54 est31 well, at least you confirm the theory xD
10:54 svenstaro can you override all the minetest optimization flags while trying with gcc?
10:54 svenstaro like, just -O0
10:55 est31 this bug is much better tested with the irrlicht examples
10:55 est31 they compile faster :)
10:55 svenstaro true
10:55 svenstaro but I'm busy right now, somebody go do that :P
10:58 svenstaro I very much suspect it's an optimization flags
10:58 svenstaro irrlicht optimitzes extremely aggressively
10:58 svenstaro and gcc is becoming better at removing unnecessary stuff from binaries
10:58 svenstaro so maybe irrlicht was building on undefined bevahior, strictly speaking
11:04 est31 in both arch and debian, minetest is the only package in official package sources that depends on the irrlicht package
11:05 Megaf hm, if I have gcc 4.9 and gcc 5.1 installed on my system, which one will Minetest use?
11:05 Megaf when I run cmake and make
11:06 sfan5 cmake tells you which compiler it uses
11:06 sfan5 (when you run it for the first time, that means no CMakeCache.txt file present)
11:06 Megaf -- The C compiler identification is GNU 4.9.3
11:06 Megaf ok
11:10 svenstaro Megaf: you can overwrite it by setting CXX
11:10 Megaf cmake options here is not complete http://dev.minetest.net/CMake_Options
11:10 Megaf svenstaro: I have no idea how to do that
11:12 * Megaf researches
11:16 svenstaro Megaf: delete your cmake cache
11:16 svenstaro then run CXX=g++-5.2 cmake
11:19 sfan5 svenstaro: also happens with -O0
11:19 sfan5 gcc
11:19 svenstaro meh
11:19 sfan5 Megaf: the compilers options are not specific to minetest and are in the cmake documentation, that wiki page only contains minetest specific stuff (mostly)
11:22 est31 but it might be outdated for some dependencies recently added, be warned
11:27 enesbil joined #minetest-dev
11:31 Megaf I cant compile with gcc5, I dont know what to put in CXX=, it cant find the compiler
11:31 Megaf I have llvm/clang too
11:33 Megaf gcc-5: error trying to exec 'cc1plus': execvp: No such file or directory
11:33 Megaf hm
11:35 Megaf sfan5: what do you type to use clang?
11:36 sfan5 CC=clang CXX=clang++ cmake .
11:37 Megaf ok, not my fault then
11:38 Megaf sfan5:  /usr/bin/ld: cannot find -lstdc++
11:38 sfan5 fix your system i guess
11:43 Samson1 joined #minetest-dev
11:56 selat joined #minetest-dev
12:40 Zeitgeist_ joined #minetest-dev
12:56 Megaf I guess I will stick to gcc 4.9
12:58 ElectronLibre joined #minetest-dev
13:06 Megaf_ joined #minetest-dev
13:45 H-H-H joined #minetest-dev
14:21 Siva joined #minetest-dev
14:30 damiel joined #minetest-dev
14:39 Zeitgeist_ joined #minetest-dev
14:48 Samson1 joined #minetest-dev
15:12 domtron joined #minetest-dev
15:13 hmmmm joined #minetest-dev
15:15 ElectronLibre joined #minetest-dev
15:29 proller joined #minetest-dev
16:00 domtron joined #minetest-dev
16:19 domtron joined #minetest-dev
16:50 Siva joined #minetest-dev
16:52 book` joined #minetest-dev
17:11 Samson1 joined #minetest-dev
17:13 Calinou I successfully learned Gulp, maybe minetest.net will use it in some time.
17:13 Calinou I have a really good workflow set up here: https://gist.github.com/Ca​linou/3cd3d529a6d3dabb911e
17:13 Calinou but it does not work on Windows (ENOENT error on "spawn jekyll"), even though Jekyll is installed and in PATH
17:20 Player_2 joined #minetest-dev
17:23 domtron joined #minetest-dev
17:24 msantana left #minetest-dev
17:30 domtron joined #minetest-dev
18:04 Samson1 joined #minetest-dev
18:16 Siva joined #minetest-dev
18:17 Player-2 joined #minetest-dev
18:18 Gael-de-Sailly joined #minetest-dev
18:39 H-H-H joined #minetest-dev
19:16 kaeza joined #minetest-dev
19:27 domtron joined #minetest-dev
19:49 Amaz joined #minetest-dev
20:04 Zeitgeist_ joined #minetest-dev
20:36 domtron joined #minetest-dev
20:55 Lunatrius joined #minetest-dev
21:19 ElectronLibre left #minetest-dev
21:23 damiel joined #minetest-dev
21:27 Zeitgeist_ joined #minetest-dev
21:33 damiel joined #minetest-dev
21:43 younishd joined #minetest-dev
21:55 Lunatrius joined #minetest-dev
22:08 EvergreenTree joined #minetest-dev
22:14 technics joined #minetest-dev
22:33 Samson1 joined #minetest-dev
22:55 domtron joined #minetest-dev
23:05 domtron joined #minetest-dev

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