mesa_temp_cache behavior

[mathren]

Hello!

When experimenting with non-standard neutrino cooling, I have encountered a weird and non-reproducible behavior, with different results depending on the presence of .mesa_temp_cache or not. This even if the .mesa_temp_cache is empty.

While digging further, I got confused on how this temporary cache actually works.

A possible brute-force fix could be adding rm -r .mesa_temp_cache to the ./clean script, but probably best to first understand why that folder gets created and how it is used.

[mathren]

Note that some module level ./clean scripts include already a removal of the cache:

rg ".mesa_temp_cache" $MESA_DIR/
[...]
./eos/eosFreeEOS_builder/clean
4:rm -rf .mesa_temp_cache

There is also a variable use_mesa_temp_cache defined in const/public/const_def.f90, which appears to be initialized to .true. in the function do_const_init in the same file. The function const_init provides a publicly available wrapper, which is then called in many places. It therefore seems like this is always set to true, but it's still not clear to me when it's actually used.