I’m not sure it’s that big of a deal; in other places where an Int is expected, as long as there’s an exact integer representation, conversion is implicit and expected:
julia> struct Foo x::Int end
julia> Foo(2e9)
Foo(2000000000)
Float64s have a 52-bit mantissa (good for about 15 sigfigs), and Int64s have a maximum value of 2^{63}-1 \approx 9.2\times 10^{18}, so there’s a large practical range of overlap.
I’m not sure I understand the question. It would be of benefit every time you need a big integer (one common one being for i in 1:2e9).
It’s not a real suggestion, though, I just miss a convenient syntax for large Ints.
I’m mostly just pondering if your desire for a convenience syntax for large Ints, couldn’t be satisfied by a few implicit Float64→Int conversions in the language.
for i=1:2e9 is a pretty good example of a place where it wouldn’t work though.
Maybe it would work, but it would be very nice, I fear.