[110106] trunk/dports/math/octave-devel/Portfile

[Michael Dickens]

On Wed, Aug 28, 2013, at 12:46 PM, Joshua Root wrote:
> The problem was defining each variant multiple times. Adding an eval to
> each of the multiple definitions wouldn't have helped.

OK; fair enough; could be that's the case. I know when I redid the
octave-devel stuff before the latest "Fortran recipe" was resolved, the
way I was able to get the variants to work was to do an "eval [subst {".
 Maybe this was not the only way, but it's what is done in other ports
for Python variants and so I did the old cut-and-paste trick which
worked quite well. Then, the "eval" shows up in the latest "Fortran
recipe" too ... so, natural conclusion. Either way, I'm glad the new
recipe is working. Mine for octave-devel is similar, and it works. So,
I'm leaving it alone for now except maybe moving back to just setting
the Fortran variables and letting the user select the C/C++ compiler
separately.  I'm not 100% sure this is a good idea, per discussion in
another email threads, but I'll think about it. - MLD