Add 32-bit Linux CI coverage#2585
Conversation
|
|
danyalahmed1995
force-pushed
the
ci/add-32bit-linux
branch
from
1e83ec5 to
8122b80
Compare
M2-TE
left a comment
M2-TE
left a comment
There was a problem hiding this comment.
Good addition, thanks for the contribution!
Just got a few comments regarding the addition of new options in the matrix entry.
|
I would also recommend using Clang with |
Co-authored-by: Jan Kuhlmann <33833587+M2-TE@users.noreply.github.com>
|
@sharadhr I went with But if you want the CI entry to model cross-compilation more explicitly, I can switch it to Clang with |
danyalahmed1995
force-pushed
the
ci/add-32bit-linux
branch
from
a7c7c23 to
5fd2c8d
Compare
Co-authored-by: Jan Kuhlmann <33833587+M2-TE@users.noreply.github.com>
danyalahmed1995
force-pushed
the
ci/add-32bit-linux
branch
from
5fd2c8d to
e8243a1
Compare
|
Looks good to me, nice work! The clang runner can be a separate PR imo. |
|
Recent force-push removed the changes made from a previous one (a7c7c23 to 5fd2c8d), so please add them back in to allow runners to succeed. Specifically, the changes made to ubuntu packages that are to be installed need to be reintroduced. Should be: sudo apt install -y libgl-dev libxcursor-dev libxi-dev libxinerama-dev libxrandr-dev ${{ contains(matrix.env.arch, 'x86') && format('{0}-multilib {1}-multilib', matrix.env.c, matrix.env.cxx) || '' }}Force pushes are evil! ;) |
|
Should be good now |
|
Thanks @danyalahmed1995 and @M2-TE for your contribution, adding this CI step! |
5 participants
Adds a 32-bit Ubuntu CI matrix entry using gcc/g++ multilib and -m32 flags.
The sample build is skipped for this cross-compilation entry to avoid requiring 32-bit graphics/system libraries, while header generation and the unit test builds still run across the supported C++ standards.
This covers the Linux 32-bit CI case discussed in #2584 and helps exercise the kind of build configuration involved in #2583.
Tested on my fork:
Refs #2584.