RE: [boost] BCB5, the pimpl idiom bug and the new Unit Test Framework

From: Fernando Cacciola (fcacciola_at_[hidden])
Date: 2001-12-21 10:21:54

----- Original Message -----
From: Peter Dimov <pdimov_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Friday, December 21, 2001 12:04 PM
Subject: Re: [boost] BCB5, the pimpl idiom bug and the new Unit Test
Framework

> From: "Fernando Cacciola" <fcacciola_at_[hidden]>
> > How about what Beman suggested about including the current impl_ptr<> as
> an
> > implementation detail inside the Test library?
> > Do you think there could be some dark corners in impl_ptr<> that could
> cause
> > problems if used in this particular way?
>
> Dark corners? In my code? Never! ;-)
>
I meant undocumented *features* :-)

> Now seriously, I don't expect that impl_ptr could cause problems. Still,
if
> I was the maintainer of Boost.Test, I wouldn't have used it. Every bit of
> additional complexity increases the probability of something going wrong
on
> a platform that the author cannot test on. (*)
>
Good point.

> Adding a detail::impl_ptr (through the back door) as a part of Boost.Test
> doesn't feel right, either; it might need to evolve a bit further before
> being ready.
>
Agreed. It doesn't feel right.

> Of course if Gennadiy decides to use impl_ptr in Boost.Test, I'll do
> whatever needs to be done.
>
Well, using plain old raw pointers won't be thread safe, but safer in any
other respect...
So I think that you are right, by now, better stick to raw pointers....

Fernando Cacciola
Sierra s.r.l.
fcacciola_at_[hidden]
www.gosierra.com