Description of problem: After installing python3 into the buildroot to work around bug 1331789, I get these errors when building cp2k. I'm not sure if they're related, but they seem to come both from scripts provided by rpm-mpi-hooks and Lmod. Version-Release number of selected component (if applicable): rpm-mpi-hooks-3-6.fc24.noarch Lmod-6.3.1-1.fc25.x86_64 How reproducible: Always. Steps to Reproduce: 1. Build current cp2k package from git master HEAD with mock -r fedora-rawhide-x86_64 Actual results: When rpmbuild gets to extracting Provides/Requires, I see tons of the errors below, repeating over and over: [...] /usr/bin/lua: /usr/share/lmod/lmod/libexec/../tools/TermWidth.lua:69: bad argument #2 to 'tonumber' (number expected, got boolean) stack traceback: [C]: in function 'tonumber' /usr/share/lmod/lmod/libexec/../tools/TermWidth.lua:69: in upvalue 'askSystem' /usr/share/lmod/lmod/libexec/../tools/TermWidth.lua:92: in function 'TermWidth' /usr/share/lmod/lmod/libexec/../tools/Banner.lua:54: in upvalue 'new' /usr/share/lmod/lmod/libexec/../tools/Banner.lua:65: in function 'Banner.banner' /usr/share/lmod/lmod/libexec/lmod:555: in function 'main' /usr/share/lmod/lmod/libexec/lmod:711: in main chunk [C]: in ? /usr/share/lmod/lmod/init/sh: line 14: 80: command not found /usr/lib/rpm/mpi.req: line 57: mpi_compiler_for_dir[${MPI_HOME}]: bad array subscript [...] Expected results: No errors during build.
Please ignore the errors from lmod - that's a separate bug that can be worked around by (for example) setting the COLUMNS environment variable: export COLUMNS=80
Looks like ${MPI_HOME} is empty?
I don't see any errors with rpm-mpi-hooks-4-1.fc25, assuming fixed.