Subranges of integers

[certik]

Extracted from #49. Author: Van Snyder

Reference: 04-140r1.

Unsigned integers, and integers that have nondecimal range, are frequent requests. These
requests could be satisfied, and numerous other benefits provided as well, by providing for
named subranges of integers.

One of the additional benefits is that checking of subscript bounds can be done at compile
time: If the bounds of an array are specified by a subrange name, and if in that case the only
subscripts are of that subrange, one only need check when the subscript gets a value, not when
it’s used as a subscript. If DO control is defined by reference to the subrange, say by using
TINY and HUGE, even that check can frequently be avoided.

Another benefit is that one can get most (or perhaps all) of the desired functionality of a BIT
type by specifying a constant subrange of 0:1.

[gronki]

Unsigned integers -- very important IMHO. Dominik

[certik]

@gronki Specifically for unsigned integers, we have a dedicated issue at #2.

[swpoole]

The lack of unsigned integers has for some communities, caused a migration to C/C++ or other more modern languages. Having unsigned_integer capabilities would be extremely useful in a number of computational venues.