Some of the ArrayFire constructors that accept the dim_t type are not
marked explicit. this allows the initialization of the ArrayFire's array
using integer types.

Description

Some of the ArrayFire constructors that accept the dim_t type are not
marked explicit. this allows the initialization of the ArrayFire's array
using integer types. For example

af::array a = 5

will create an af::array with 5 elements. This is not intended behavior.

I have looked into the ABI for this change and it doesn't seem to
be affected on GCC. I have to still test this on MSVC. This CAN
break some existing code because it does change the API but ArrayFire
was never designed with this code in mind.

Changes to Users

This change CAN break some user code. The following code will break

• Initialize af::array using an integer or float

array a = 5;
array b = 5.0;

• Pass an integer to a function that accepts an array

void func(af::array a) {
    ...
}

func(5);

Checklist

• Rebased on latest master
• Code compiles
• Tests pass
• [ ] Functions added to unified API
• [ ] Functions documented