|Copyright||(c) The University of Glasgow 2001|
|License||BSD-style (see the file libraries/base/LICENSE)|
Unsigned integer types.
Unsigned integral types
8-bit unsigned integer type
16-bit unsigned integer type
32-bit unsigned integer type
64-bit unsigned integer type
- All arithmetic is performed modulo 2^n, where n is the number of
bits in the type. One non-obvious consequence of this is that
negateshould not raise an error on negative arguments.
- For coercing between any two integer types, use
fromIntegral, which is specialized for all the common cases so should be fast enough. Coercing word types to and from integer types preserves representation, not sign.
- It would be very natural to add a type
Naturalproviding an unbounded size unsigned integer, just as
Integerprovides unbounded size signed integers. We do not do that yet since there is no demand for it.
- The rules that hold for
Enuminstances over a bounded type such as
Int(see the section of the Haskell report dealing with arithmetic sequences) also hold for the
Enuminstances over the various
Wordtypes defined here.
- Right and left shifts by amounts greater than or equal to the width
of the type result in a zero result. This is contrary to the
behaviour in C, which is undefined; a common interpretation is to
truncate the shift count to the width of the type, for example
1 << 32 == 1in some C implementations.