[127652] trunk/dports/python

[Ryan Schmidt]

> On Oct 31, 2014, at 9:57 AM, stromnov at macports.org wrote:
> 
> Revision
> 127652
> Author
> stromnov at macports.org
> Date
> 2014-10-31 07:57:55 -0700 (Fri, 31 Oct 2014)
> Log Message
> 
> py-networkit: new port

> Added: trunk/dports/python/py-networkit/Portfile (0 => 127652)

> +    variant gcc48 conflicts gcc49 description {build using macports-gcc-4.8} {}
> +    variant gcc49 conflicts gcc48 description {build using macports-gcc-4.9} {}
> +
> +    if {![variant_isset gcc48] && ![variant_isset gcc49]} {
> +        default_variants +gcc48
> +    }
> +
> +    if {[variant_isset gcc48]} {
> +       configure.compiler   macports-gcc-4.8
> +    } elseif {[variant_isset gcc49]} {
> +       configure.compiler   macports-gcc-4.9
> +    }

Is there a reason why this configure.compiler directive isn't being placed inside the variant block itself, like we normally do? That would be clearer:


    variant gcc48 conflicts gcc49 description {build using macports-gcc-4.8} {
        configure.compiler   macports-gcc-4.8
    }

    variant gcc49 conflicts gcc48 description {build using macports-gcc-4.9} {
        configure.compiler   macports-gcc-4.9
    }

    if {![variant_isset gcc48] && ![variant_isset gcc49]} {
        default_variants +gcc48
    }


Also, what is this port's special reliance on gcc that clang would not work?