Re: [boost] (RFC) make_shared name change

From: Peter Dimov (pdimov_at_[hidden])
Date: 2003-01-14 12:29:04

From: "Daniel Frey" <daniel.frey_at_[hidden]>
> Peter Dimov wrote:
> >
> > Perhaps, but some names are "less bad". It is a convention that make_*
names
> > are constructor aliases; this is not the case here, so I conclude that
> > make_shared isn't a particularly good choice.
>
> I've always seen it another way: make_* names are convenience functions
> which use argument deduction to save me some typing (there are even
> cases where I can't write the type down easily).

Good summary, I agree.

> I don't expect them to
> match on constructors directly. Is there any reference about which is
> the "right" meaning of make_*? What do others think?

FWIW, I've seen people incorrectly deduce from the name that make_shared has
the same semantics as the constructor.

> > >From semantic point of view, it's more like "This function will obtain
a
> > shared_ptr to the object identified by the weak_ptr argument. If there
is no
> > such object, an empty shared_ptr is returned."
>
> Which omits the type deduction part. It seems that this is why I have
> problems moving away from make_*...

That's true. The intent of make_shared/get_shared_ptr has nothing to do with
argument deduction, and it doesn't save any typing:

if(shared_ptr<X> px = make_shared(wp))
{
// ...
}

You still need to spell "shared_ptr<X>" in its entirety. The idiomatic use
doesn't match the name.