lowflow doesn't compile due to rank mismatch

[Peter Hill]

lowflow.f90:220:27:

                 w_r(ie) = w(ie) * exp(energy(ie)*(1.0-1.0/temp_neo(ir,is))) / temp_neo(ir,is)**1.5
                           1
Error: Rank mismatch in array reference at (1) (1/2)

w is w(negrid,nspec).

As a secondary bug, w doesn't appear to be documented.

[Joseph Parker]

Thanks Peter. This dates to the non-Maxwellian work, which updated things like energy and w to be functions of is as well as ie. (The exact commit is 92896fc.) This could easily be missed, because lowflow has to be explicitly turned on at compile time.

[Peter Hill]

It might be wise to compile on Jenkins with all the optional features turned on. Are there any features that conflict? We might want at least MPI/non-MPI builds

[David Dickinson]

@ZedThree I'm not aware of any conflicting features.

[Joseph Parker]

Me neither. I'm also not aware of non-MPI users! ;)

It didn't auto-comment, but this is addressed by PR #33

[David Dickinson]

@josephparker compiling without mpi can be very helpful for some types of debugging (e.g. to avoid false positives in valgrind).

[Joseph Parker]

See also #37

[Joseph Parker]

Resolved by PR #33