[Boost-users] Re: Inserting elements in extensible associative sequences

From: David Abrahams (dave_at_[hidden])
Date: 2005-02-15 13:51:18

"Lemings, Eric B. \(Contr\)" <Eric.Lemings_at_[hidden]> writes:

> The semantics of MPL's insert metafunction for an extensible associative
> sequence is defined as:
>
> "For any Extensible Associative Sequence s, iterator pos in s, and
> arbitrary type x:
>
> typedef insert<s,x>::type r;
>
> Semantics: r is concept-identical and equivalent to s, except that
> at< r, key_type<s,x>::type >::type is identical to value_type<s,x>::type."
>
> What happens if the key type of x is already in s?

That's covered by the semantics; it's not a special case.  The new
value value_type<s,x>::type is associated with the key
key_type<s,x>::type in r.

HTH,

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com