Fix incorrect string manipulation

[wrwilliams]

rstrip was of course intended, not sure how remove_suffix worked under any conditions.

Fixes #43716.

[wspear]

This still doesn't work for me. I get: Error: AttributeError: type object 'super' has no attribute 'with_or_without'

[wspear]

This fails like:
==> Error: AttributeError: 'list' object has no attribute 'replace'

Because with_or_without returns a list of arguments, not a single string. Are you able to test these changes against spack develop before committing? Do you have a test case where the "=yes" string is present? It doesn't show up when I print the results of with_or_without from my default scorep install. (The only argument returned for me is "--without-shmem")

[wrwilliams]

This fails like: ==> Error: AttributeError: 'list' object has no attribute 'replace'

Because with_or_without returns a list of arguments, not a single string. Are you able to test these changes against spack develop before committing? Do you have a test case where the "=yes" string is present? It doesn't show up when I print the results of with_or_without from my default scorep install. (The only argument returned for me is "--without-shmem")

I do not have a working develop environment at present, and do not have many cycles free to create one. A working test case for the =yes string should be generated via any MPI or SHMEM variant that is not explicitly included in the subsequent mapping (e.g. OneAPI MPI, see #43700).

--with-mpi should work on ~99% of systems as long as a) there is only one MPI present at build time, which Spack should ensure is the case, and b) the MPI in question doesn't do anything weird that keeps us from autodetecting it as OMPI/MPICH/Intel. --with-mpi=yes (ditto --with-shmem=yes) will simply not work.

If --with-or-without cannot handle not having a with-value, it is the wrong tool and the package should instead map any dependency on the virtual MPI target to --with-mpi manually.

[wspear]

I wasn't sure if this should go to the issue or this PR... but for reference: #43700 (comment)