Current Server Time: 23:09 (Central Europe)

#apertus IRC Channel Logs

2019/10/29

Timezone: UTC


00:14
Spirit532
left the channel
00:26
Spirit532
joined the channel
02:34
futarisIRCcloud
joined the channel
02:55
intrac
changed nick to: intracube
07:48
aombk2
left the channel
07:56
aombk
joined the channel
08:09
Bertl_zZ
changed nick to: Bertl
08:09
Bertl
morning folks!
08:26
sebix
joined the channel
08:26
sebix
left the channel
08:26
sebix
joined the channel
10:20
illwieckz
left the channel
13:53
illwieckz
joined the channel
14:44
Bertl
off for now .. bbl
14:44
Bertl
changed nick to: Bertl_oO
14:58
BAndiT1983|away
changed nick to: BAndiT1983
15:07
intracube
changed nick to: intrac
15:07
intrac
changed nick to: intracube
15:59
RexOrCine|away
changed nick to: RexOrCine
16:15
BAndiT1983
changed nick to: BAndiT1983|away
16:49
sebix
left the channel
17:43
BAndiT1983|away
changed nick to: BAndiT1983
20:31
BAndiT1983
anuejn, are you there?
20:45
vup
i am here
20:47
BAndiT1983
ah, sorry constantly mixing you both up ;)
20:47
vup
no worries
20:47
BAndiT1983
have commented on the environment var for C++ compiler, somehow it is missing
20:47
vup
yes
20:47
vup
that was also one of my first guesse, but later on it says
20:47
vup
-- Check for working C compiler: /usr/sbin/cc -- works
20:48
vup
why just doesn't make any sense to me
20:49
BAndiT1983
your line is about C compiler, but i see in the logs, that it finds c++ stuff, although i would'Ve expected gcc or clang
20:49
BAndiT1983
are cc and c++ aliases?
20:50
vup
it also says
20:50
vup
-- Check for working CXX compiler: /usr/sbin/c++ -- works
20:50
vup
(and of course -- The CXX compiler identification is unknown)
20:50
BAndiT1983
yup
20:50
BAndiT1983
hope that this is not a bug in cmake 3.15, as this link talks about a fix for 3.16 -> https://stackoverflow.com/questions/56937736/ti-cl2000-cmake-cxx-compiler-identification-is-unknown
20:51
BAndiT1983
would it help to supply -DCMAKE_CXX_COMPILER=... var for now?
20:51
vup
yes atleast locally /usr/sbin/cc is a link to gcc
20:51
BAndiT1983
i know that the link is talking about different compiler, but the bug just occurred recently it seems
20:52
vup
/usr/sbin/c++ is a executable, but seems too small to be the actual compiler
20:52
vup
i can try -DCMAKE_CXX_COMPILER=
20:52
vup
in the next CI run, but the error seems different here
20:52
BAndiT1983
is it a symlink?
20:52
vup
the c++ executable?
20:52
vup
no
20:53
BAndiT1983
i would almost guess that the compiler is not installed correctly
20:53
vup
hmm
20:53
futarisIRCcloud
left the channel
20:54
vup
ok seems it is either a hardlink or a copy
20:54
vup
but it has the same checksum as g++
20:54
vup
so that seems fine?
20:54
BAndiT1983
could be
20:54
BAndiT1983
is your container up to date?
20:54
vup
this is not even in the docker container
20:55
vup
but in the archlinuxarm chroot
20:55
BAndiT1983
ah, ok
20:55
vup
which should be exactly the same
20:55
vup
(minus updates between building locally and in the CI)
20:56
vup
(but updates can't be the problem, as i built it locally successfully after it broke in the CI)
20:57
BAndiT1983
just scrolled through the log and i see several package installs, but don't know the sequence of steps
20:57
BAndiT1983
wondering about gcc and clang install, but normally it shouldn't collide
20:57
vup
yeah and it works when building locally
20:58
BAndiT1983
CMake Error at /usr/share/cmake-3.15/Modules/CMakeCompilerIdDetection.cmake:26 (list):
20:58
BAndiT1983
can't we use 3.14 or so, just to check if there is a bug?
20:58
vup
the whole thing works fine if i run it locally
20:59
vup
so cmake 3.15 shouldn't be the problem
20:59
BAndiT1983
ok,ok, just looking for clues, as we had CI problems previously, which was working locally, but not on the build machine
20:59
vup
i just don't understand how the CI system can be any different from local
21:01
futarisIRCcloud
joined the channel
21:01
BAndiT1983
had trouble with some of the layers CI system use above containers or so, it wasn't building correctly and required a bit of voodoo, which differed from local one just slightly
21:01
vup
we are building everything in a docker container, and this doesn't even fail in the docker container, but in a chroot in this docker container, which again shouldn't have any differences between local and CI
21:01
BAndiT1983
is it possible to run CirrusCI locally? was testing my build scripts for CircleCI that way
21:01
vup
nope can't run cirrusci stuff locally yet
21:02
BAndiT1983
hm, that's a pity, this would rule out any possible intervention in their system
21:02
vup
yeah
21:02
vup
but i also don't quite see a way yet the actually could interfer
21:03
vup
hmm let me try to update the build container
21:04
BAndiT1983
it was quite some time ago, but there were some differences between local docker container and what CircleCI (or TravisCI?) have done while building
21:05
BAndiT1983
so it was pure luck, that it was possible to execute things on local machine to locate the problem
21:05
vup
hmm strange
21:05
BAndiT1983
e.g. if you don't install git in the container, then it uses it's own isntallation, from wherever
21:06
vup
hmm
21:06
vup
we are manually calling docker in the CI
21:06
BAndiT1983
but as i required submodules, i've added git as package
21:06
BAndiT1983
ok, then it's really interesting what's going on
21:06
vup
yep
21:06
BAndiT1983
have you built with version infos yet?
21:08
vup
yeah
21:08
vup
https://paste.niemo.de/mekekuvevi
21:08
BAndiT1983
looks also normal
21:08
vup
indeed
21:09
BAndiT1983
is g++ available as command?
21:10
BAndiT1983
try "which g++"
21:10
BAndiT1983
also "which c++", but just to see if it is found correctly
21:12
vup
sure
21:12
vup
one moment
22:33
BAndiT1983
changed nick to: BAndiT1983|away
23:03
futarisIRCcloud
left the channel
23:05
futarisIRCcloud
joined the channel