cvs commit: ports/astro/nightfall Makefile ports/benchmarks/himenobench Makefile ports/benchmarks/hpl Makefile ports/biology/molden Makefile ports/biology/ortep3 Makefile ports/biology/platon Makefile ports/biology/psi88 Makefile ports/biology/tinker ...

[Gerald Pfeifer]

On Thu, 8 Jan 2009, Pav Lucistnik wrote:
> Because all the fortran code must be recompiled against new and
> incompatible fortran library.

Yep.  Strictly speaking, only ports with USE_FORTRAN=yes that depend on 
some other port that has USE_FORTRAN=yes, or ports with USE_FORTRAN=yes 
which some  other port that has USE_FORTRAN=yes depends on would have
needed the bump (modulo potential additional "manual" dependencies).

I believe we do not have a reasonable way to determine such a cover, and
all those OPTIONS and knobs make this even more tricky.  And the result 
likely would have been pretty close to the full set anyway...

Gerald
-- 
Gerald (Jerry) Pfeifer   gerald at pfeifer.com   http://www.pfeifer.com/gerald/