Date: 2001-06-08 14:11:36
--- In boost_at_y..., Mike Gradman <windryder1_at_y...> wrote:
> The real solution to the problem shouldn't be handled
> in a library given all of these potential #define
> problems ... the best thing might be an actual
> language extension to support the desired definition
> of NULL ... that would have to go through the
> standards process ... in the meantime, why not just
> use plain old 0?
You miss the reason for having a "nil". A nil value has a specific
type, nil_t, which can be used in ways that are safer than using a
0. Read the full thread on this to understand.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk