compilers-1.0 port group

[Sean Farley]

Hi all,

This is a follow-up to my request for comments started here:

https://lists.macosforge.org/pipermail/macports-dev/2013-July/023410.html

I've made a lot of improvements based on the feedback I got (thanks!)
and would like to try to get these port groups into trunk so I can
finally stop putting off my study work involving mpi ;-)

Here is the newest version:

https://bitbucket.org/seanfarley/scienceports/src/808f6c65589ebc290ef913c0c6e2cca580026110/_resources/port1.0/group/compilers-1.0.tcl?at=default

but I'll point out some features here:

* compilers.setup takes blacklisted or whitelisted variants, e.g. "-gcc
  +gcc45" would blacklist all gcc compilers but whitelist gcc45
* if no arguments are given to compilers.setup then all variants are
  whitelisted
* there is an optional "require_fortran" argument to send
  compilers.setup to specify that a fortran compiler is needed
* there is a helper procedure called "fortran_variant_isset" to help
  port authors write special logic if fortran is specified
* similar for the procedure "fortran_variant_name"

As a bonus, I believe fortran recipes like the one Jeremy sent earlier
today (what a coincidence)

https://lists.macosforge.org/pipermail/macports-dev/2013-August/023908.html

could all be unified as such:

{{{
PortGroup compilers 1.0

compilers.setup     -clang require_fortran
}}}

If this passes review, then I'll start another thread about mpi port
group which, for the curious, is here:

https://bitbucket.org/seanfarley/scienceports/src/808f6c65589ebc290ef913c0c6e2cca580026110/_resources/port1.0/group/mpi-1.0.tcl?at=default