|
Boost : |
From: David Abrahams (david.abrahams_at_[hidden])
Date: 2001-07-24 09:47:06
----- Original Message -----
From: "Jeremy Siek" <jsiek_at_[hidden]>
> The property_num business was a workaround for not having a version of
> is_same working for compilers without partial specialization... taking a
> quick look at type_traits, we now have a version of is_same that doesn't
> use partial spec., so we should be able to ditch that.
Great!
BTW, I think I understand what the "property" or "compile-time associative
container" object should be called, now: I think it is a "tagged tuple". And
I think it would be best implemented as a tuple of tagged_value<T> objects:
template <class Tag, class Value>
struct tagged_value
{
typedef Tag tag;
typedef Value value_type;
tagged_value(const Value& v) : value(v) {}
value_type value;
};
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk