ci debugging

[lmmx]

Let's goooo

• linux
  • ubuntu-22.04 x86_64
  • ubuntu-22.04 i686-unknown-linux-gnu
  • ubuntu-22.04 aarch64
  • ubuntu-22.04 armv7
  • ubuntu-22.04 s390x
  • ubuntu-22.04 ppc64le
• musllinux
  • ubuntu-22.04 x86_64
  • ubuntu-22.04 i686
  • ubuntu-22.04 aarch64
  • ubuntu-22.04 armv7
• windows
  • windows-latest x64
  • windows-latest x86
• macos
  • macos-13 x86_64
  • macos-14 aarch64

[lmmx]

aarch built! 🎉

but... it seems to have regressed i686 in the process (killed the job after 2 hours, perhaps endless build, even if not that is not acceptable really)

[lmmx]

merge commit messed up the lock file

[lmmx]

Action won't run because of its own bad merge messing up the lock file

[lmmx]

i686 broke but now due to linker failures

It expects fortran and python 3.9

   = note: /opt/rh/gcc-toolset-14/root/usr/libexec/gcc/x86_64-redhat-linux/14/ld: skipping incompatible /opt/rh/gcc-toolset-14/root/usr/lib/gcc/x86_64-redhat-linux/14/libgfortran_nonshared.a when searching for -lgfortran_nonshared
          /opt/rh/gcc-toolset-14/root/usr/libexec/gcc/x86_64-redhat-linux/14/ld: cannot find -lgfortran_nonshared: No such file or directory
          /opt/rh/gcc-toolset-14/root/usr/libexec/gcc/x86_64-redhat-linux/14/ld: cannot find -lpython3.9: No such file or directory

[lmmx]

could try turning off the openblas system feature for aarch, that seemed to work OK when I accidentally wasn't passing the features at all (took 9 mins), it was just the i686 that wouldn't work