Re: [boost] [optional] Changes in Boost.Optional

Subject: Re: [boost] [optional] Changes in Boost.Optional
From: Andrzej Krzemienski (akrzemi1_at_[hidden])
Date: 2014-11-24 08:53:16

2014-09-02 16:11 GMT+02:00 Dean Michael Berris <mikhailberis_at_[hidden]>:

> On Wed Sep 03 2014 at 12:00:56 AM Andrzej Krzemienski <akrzemi1_at_[hidden]>
> wrote:
>
> >
> > And the above incorrect behaviour is the consequence of a more simple bug
> > in VC++. It allows two user defined conversions in copy initialization:
> >
> > struct A
> > {
> >   A(int &&) {}
> >
> > };
> >
> > struct B
> > {
> >   operator int() { return 0; }
> > };
> >
> > int main()
> > {
> >   A t = B();
> > }
> >
> > This works in VC++, although the code is incorrect.
> >
> >
> So it does sound like this is one of those r-value reference implementation
> bugs in MSVC that I've heard before but couldn't cite. Could it just be
> that Boost.Optional shouldn't turn on rvalue reference support with MSVC
> 2010? Will defining just the copy constructor "fix" it? More importantly is
> there something users of Boost 1.56.0 can do to side-step this particular
> issue in Boost.Optional?
>

Would it work for you, if I added an option to configure Boost.Optional at
compile time so that it does not have these constructors:

optional(T const&);
optional(T &&);

but instead has these:

optional(ConvertibleFromTConstRefWrapper);
optional(ConvertibleFromTRefRefWrapper);

This would fix your problem on MSVC, remove all double-conversions to
optional and provide correct single-conversions by turning MSVC's
double-conversions. It wouldn't be enabled by default of course. MSVC users
are used to double conversions. I often found myself writing:

optional<std::string> os = "c-style string";

Regards,
&rzej