ks_spectrum test comparisons use different input file formats

[smangham]

I'm attempting to build and test milc, but can't really tell if it's working as the tests appear to be based on a much older version of the code.

Building ks_spectrum and running make check "EXEC=ks_spectrum_hisq" &> make_test.log, all the tests fail. As an example, ks_spectrum_hisq.nd1.test-out ends up with:

get_i: ERROR IN INPUT: expected max_cg_iterations but found set_type
exit: Thu Jul 18 16:27:06 2024

Comparing the generated output and the sample output, it looks like there's substantial differences:

It looks like these arise at least in part because the input file ks_spectrum_hisq.nd1.test-out does not actually reflect the file that's supposed to be built by build_input.ks_spectrum_hisq.nd.1.sh - the sample.in file has a set_type multimass line that build_input no longer uses. I'm not sure which command or part of the make chain would actually force a rebuild of those sample.in files.

I'll try just deleting them to force a rebuild, but I think they should probably be either removed from the repo, or the makefiles should be adjusted to force an update.

[smangham]

Right, I think I had it back to front - deleting the set_type multimass line from the sample-in files causes the tests to pass.

[detar]

I hope you are using the "develop" branch.