[cfp-interest 3747] INFINITY on a system supporting _Float16 - not urgent

[Damian McGuckin]

What is the semantics of a comparison of INFINITY against a _Float16?

Prior to _Float16, a comparison of a real floating type against INFINITY 
would never have touched the real floating type but just converted the
INFINITY to the appropriate type (likely at compile) because INFINITY was 
defined as a float (the smallest of the real floating types at the time).

That all changes with _Float16.

Thanks - Damian