Refactor read_extra_..._inlist into arrays

[warrickball]

Since time immemorial (before r4366), various MESA inlists have had the option to read other inlists recursively, with options like

    read_extra_star_job_inlist1 = .false.
    extra_star_job_inlist1_name = 'undefined'
    read_extra_star_job_inlist2 = .false.
    extra_star_job_inlist2_name = 'undefined'
    read_extra_star_job_inlist3 = .false.
    extra_star_job_inlist3_name = 'undefined'
    read_extra_star_job_inlist4 = .false.
    extra_star_job_inlist4_name = 'undefined'
    read_extra_star_job_inlist5 = .false.
    extra_star_job_inlist5_name = 'undefined'

The canonical default inlist itself uses this to load other inlists.

Should we change these to arrays? I say yes. We can also scrap the logical flag and just use the empty string '' to indicate that an inlist should not be read (unless I'm missing a use case where this helpful). e.g. I imagine a user would write

    extra_star_job_inlist_name(1) = 'inlist_project'

and the defaults file containing

    extra_star_job_inlist_name(:) = ''

to cover all values that aren't used. E.g., in a loop like this should work:

do i=1, max_extra_inlists
   if (extra_star_job_inlist_name(i) == '') continue
   ... ! otherwise read it

If the logical flag is useful, we can instead use

do i=1, max_extra_inlists
   if (.not. read_extra_star_job_inlist(i)) continue
   ... ! otherwise read it

The only downside I can think of is that this will (trivially) break some backwards compatibility but I don't think that's a strong argument.

[fxt44]

let me ask, what is gained by moving to arrays?

i suspect most of the test suite will need adjustments to make this work ...

[warrickball]

let me ask, what is gained by moving to arrays?

We can refactor processing the inlists into loops, which at a rough count would eliminate ~300 lines of code. It also means that future changes only have to be applied once, not 5 times. Plus (though I don't think anyone has ever asked for more) it becomes trivial to increase the limit past 5.

i suspect most of the test suite will need adjustments to make this work ...

I think a few tactical seds will take care of almost all of these.

[fxt44]

fine by me. sharing said tactical sed scripts might be useful to some users.