Add idiomatic C++ functions in bezier namespace

[dhermes]

For example, helpers.h would become

#ifndef BEZIER_HELPERS_H
#define BEZIER_HELPERS_H

#include <stdbool.h>

#if defined(__cplusplus)
extern "C" {
#endif

void BEZ_cross_product(const double* vec0, const double* vec1, double* result);
...

#if defined(__cplusplus)
}
#endif

// BEGIN: New definitions
#if defined(__cplusplus)
#include <vector>

namespace bezier {
    double cross_product(const std::vector<double>& vec0, const std::vector<double>& vec1) {
        double result;
        BEZ_cross_product(vec0.data(), vec1.data(), &result);
        return result;
    }
    ...
}
#endif
//   END: New definitions

#endif /* BEZIER_HELPERS_H */

My biggest concern is about the level of effort needed to make an idiomatic C++ interface (e.g. using vector<double> instead of double*) and using RAII.

FWIW I'm not 100% sure how I'll get the C++ name mangling to work right with the symbols exported by Fortran.

See also: #166.

[dhermes]

UPDATE: The below is outdated, the main issue description now references this core idea (after #166 was completed).

---

From a StackOverflow answer:

namespace bezier {
    extern "C" {
        void cross_product(double* vec0, double* vec1, double* result);
        ...
    }
}

is possible, though if we want the C / ABI BEZ_ prefix, it'd probably have to instead be

extern "C" {
    void BEZ_cross_product(double* vec0, double* vec1, double* result);
    ...
}

namespace bezier {
    void cross_product(double* vec0, double* vec1, double* result) {
        BEZ_cross_product(double* vec0, double* vec1, double* result);
    }
    ...
}

cc @pazner

[dhermes]

Did some Googling and found a few things:

• I think this will be pretty useful
• https://github.com/ishbir/cppnats may help as a reference point

[dhermes]

I was curious about the overhead of the vector<double> example: https://godbolt.org/z/SkLvqF

bezier::cross_product(std::vector<double, std::allocator<double> > const&, std::vector<double, std::allocator<double> > const&):
        sub     rsp, 24
        mov     rsi, QWORD PTR [rsi]
        mov     rdi, QWORD PTR [rdi]
        lea     rdx, [rsp+8]
        call    BEZ_cross_product
        movsd   xmm0, QWORD PTR [rsp+8]
        add     rsp, 24
        ret