Re: [boost] shared_array.size()

From: David Abrahams (abrahams_at_[hidden])
Date: 2000-12-13 07:56:11

Do you worry about these differences?
    vector<int> x(0);   // ok
    vector<int> x(12); // ok
and
    boost::shared_array<int> x(0);   // ok
    boost::shared_array<int> x(12)  // error

-Dave
----- Original Message -----
From: "Greg Colvin" <gcolvin_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Wednesday, December 13, 2000 12:41 AM
Subject: Re: [boost] shared_array.size()

> It's the name that worries me, it's the differing semantics,
> I think null pointer to array versus pointer to empty array?
>
> From: "David Abrahams" <abrahams_at_[hidden]>
> > "shared_array2" was just a placeholder
> >
> > any-resemblance-to-class-temlates-living-or-dead-is-pure-coincidence-ly
> > y'rs,
> > dave
> > ----- Original Message -----
> > From: "Greg Colvin" <gcolvin_at_[hidden]>
> > To: <boost_at_[hidden]>
> > Sent: Tuesday, December 12, 2000 11:46 PM
> > Subject: Re: [boost] shared_array.size()
> >
> >
> > > Yep, and adding a size field to shared_array is extra overhead as
well,
> > > but maybe not so bad.  What I don't like is setting the size field
> > > separately, but I'm not sure I like your shared_array2 interface
either.
> > > In particular, what about
> > >
> > >    boost::shared_array<int> x(0);
> > >
> > > vs.
> > >
> > >    boost::shared_array2<int> x(0);
> > >
> > > From: "David Abrahams" <abrahams_at_[hidden]>
> > > > Okay; there's one extra level of indirection in that case, though,
and a
> > bit
> > > > more overhead, FWIW.
> > > >
> > > > ----- Original Message -----
> > > > From: "Greg Colvin" <gcolvin_at_[hidden]>
> > > >
> > > >
> > > > > So use shared_ptr<vector> ?
> > > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > >
> > >
> >
> >
> >
> >
> >
>
>
>
>