28 Feb
2002
28 Feb
'02
2:04 p.m.
| By the way, the report says: "Class Enum defines operations | on sequentially | ordered types." I think the "sequentially" means that | multiple applications | of succ to a value shall not result in the same value. That's | why I think | that succ (maxBound :: Int) should be an error and not | minBound :: Int. | Because i + 1 should be equal to succ i for every Int value | i, (maxBound :: | Int) + 1 should also be an error and so should every | arithmetic operation | that overflows or underflows. The revised version of the report http://research.microsoft.com/~simonpj/haskell98-revised says that succ maxBound is an error for any Bounded type. So I think this is covered already. Simon