Idea for more delegations in <algorithm>

[Gabriel Dos Reis]

Paolo Carlini <pcarlini@suse.de> writes:

| Paolo Carlini wrote:
| 
| > Thanks for the clarification. These considerations, however, make me
| > think
| > that maybe we should *not* use __enable_if elsewhere, if ever
| > possible... ;)
| >
| > In implementing tr1/type_traits I avoided completely __enable_if,
| > but only on
| > a stylistic base, so to speak, lucky guy... ;)
| 
| ... but, thinking more about this, probably the real issue is with
| *defaults* not
| with __enable_if "per se".

Yes.  Usually, the enable_if<> stuff is put in the return type.
Constructors don't have return type, so you would have to stick it
somewhere. 

-- Gaby