non-vacuum support

[PieterTack]

Change calculation to allow for other-than-vacuum media inside the optic.

This should include the following things:

• extra parameter in polycap_description to supply medium composition
• calculation of refractive index for chosen medium, and using it in Fresnel equation (instead of using n=1)
• taking into account energy dependent absorption weighting along travel path.

Concerns:

• Do I ignore medium between source and optic? Or do I include this?

[tschoonj]

Good idea.

Perhaps we should also introduce something like polycap_composition:

typedef struct _polycap_composition              polycap_composition;
struct _polycap_composition {
    unsigned int nelem;
    int iz[];
    double wi[];
    double density;
};

This would simplify polycap_description_new to:

polycap_description* polycap_description_new(
    polycap_profile *profile,
    double sig_rough,
    int64_t n_cap,
    polycap_composition *wall_composition,
        polycap_composition *atmosphere_composition, /* NULL is acceptable here */
    polycap_error **error);

I would include the medium between source and optic in the calculations as it cannot be appropriately taken into account afterwards