Make rl.MatrixToFloatV do same thing in Odin as in C

[karl-zylinski]

Make rl.MatrixToFloatV transpose the matrix before transmuting it to [16]f32, so it does the same thing as the raymath version implemented in C.

Context:

The raymath.h version does this:

RMAPI float16 MatrixToFloatV(Matrix mat)
{
    float16 result = { 0 };

    result.v[0] = mat.m0;
    result.v[1] = mat.m1;
    result.v[2] = mat.m2;
    result.v[3] = mat.m3;
    result.v[4] = mat.m4;
    result.v[5] = mat.m5;
    result.v[6] = mat.m6;
    result.v[7] = mat.m7;
    result.v[8] = mat.m8;
    result.v[9] = mat.m9;
    result.v[10] = mat.m10;
    result.v[11] = mat.m11;
    result.v[12] = mat.m12;
    result.v[13] = mat.m13;
    result.v[14] = mat.m14;
    result.v[15] = mat.m15;

    return result;
}

But mat.m1 is actually the fifth float in it from a memory perspective, so it is equivalent to a transpose.