Re: [boost] Suggestion: scoped_ptr::release()

From: Joseph Gottman (joegottman_at_[hidden])
Date: 2000-12-27 21:08:04

All right then, how about scoped_array::release()?  auto_ptr isn't an option
in this case.

----- Original Message -----
From: "Beman Dawes" <beman_at_[hidden]>
To: <boost_at_[hidden]>; <boost_at_[hidden]>
Sent: Wednesday, December 27, 2000 12:07 PM
Subject: Re: [boost] Suggestion: scoped_ptr::release()

> At 09:23 AM 12/27/2000 -0500, David Abrahams wrote:
>
>  >I always thought that "no transfer of ownership" was part of the point
of
>  >scoped_ptr<>. Why not just use std::auto_ptr<> if you want to do this?
>
> Yes, why not?
>
> The decision not to include release() was a conscious one; basically I'd
> like to see that reaffirmed, and then I'll add a FAQ question "why isn't
> there a release()?" to the docs.
>
> --Beman
>
>  >> At 07:38 PM 12/19/2000 -0500, Joe Gottman wrote:
>  >>
>  >>      I think the scoped_ptr class should have a release() function,
>  >similar
>  >> to the one in std::auto_ptr.  This would help in cases like the
> following
>  >>
>  >> int *foo() {
>  >>      scoped_ptr<int> the_pointer(new int());
>  >>       //Do stuff that might throw
>  >>      return the_pointer.release();
>  >> }
>
>
>
>
>