New port, science/hypre

[jrm]

Summary

If CONFIGURE_TARGET is unset and GNU_CONFIGURE=yes is used, the build also works
and it builds with mpicc/mpicxx. Ignoring the faster builds and some possible
problems on other architectures, is building with mpicc/mpicxx beneficial?

Test Plan

portlint -C: OK
testport: OK (poudriere: 10.3-RELEASE-p10, i386)
testport: OK (poudriere: 10.3-RELEASE-p10, amd64)
testport: OK (poudriere: 11.0-RELEASE-p1, i386)
testport: OK (poudriere: 11.0-RELEASE-p1, amd64)

[jrm]

• Set LICENSE_FILE
• Use the RLN macro, so that fully qualified paths can be used

[pfg]

Most of the packages that use this involve MPI: I need the MPICH build for another port I am working in (sundials), however we could also use OpenMPI.

Now that I think about it, it would be nice to have an option to choose to build with OpenMPI or MPICH (MPICH still being the default).

[pfg]

science/hypre/Makefile

15 ↗

(On Diff #29200)

I think the version .8 may be needed, otherwise it may take mpich3 instead.

[jrm]

• Add knobs to choose to depend on and build with either mpich2 or openmpi

[pfg]

LGTM, Thanks!