[Development] Error "invokes deleted constructor" while building Qt5 with clang and libc++

[Francisco Lopes]

2012/12/3 Thiago Macieira <thiago.macieira at intel.com>

> On segunda-feira, 3 de dezembro de 2012 15.37.13, Francisco Lopes wrote:
> > Please take a look at my last comment on the bug report (
> > http://llvm.org/bugs/show_bug.cgi?id=14486#c10). If my analysis is
> right,
> > it's wrong to use "= {0}" with an inaccessible copy constructor.
>
> Looking at N3337, this is the API required for std::atomic:
>
> atomic() noexcept = default;
> constexpr atomic(T) noexcept;
> atomic(const atomic&) = delete;
> atomic& operator=(const atomic&) = delete;
> atomic& operator=(const atomic&) volatile = delete;
> T operator=(T) volatile noexcept;
> T operator=(T) noexcept;
>
> It also defines ATOMIC_VAR_INIT and in 29.6.5
> [atomics.types.operations.req]
> paragraph 6 we find this example:
>
> atomic<int> v = ATOMIC_VAR_INIT(5);
>
> The standard does not define what ATOMIC_VAR_INIT is. GCC defines it as:
>
> #define ATOMIC_VAR_INIT(_VI) { _VI }
>
> And clang defines it (see [1]) as:
> #define ATOMIC_VAR_INIT(__v) {__v}
>
> With constructors:
>     _LIBCPP_INLINE_VISIBILITY
>     atomic() _NOEXCEPT {} // = default;
>     _LIBCPP_INLINE_VISIBILITY
>     _LIBCPP_CONSTEXPR atomic(_Tp __d) _NOEXCEPT : __base(__d) {}
>
>     _LIBCPP_INLINE_VISIBILITY
>     _Tp operator=(_Tp __d) volatile _NOEXCEPT
>         {__base::store(__d); return __d;}
>     _LIBCPP_INLINE_VISIBILITY
>     _Tp operator=(_Tp __d) _NOEXCEPT
>         {__base::store(__d); return __d;}
>
> inheriting:
>     __atomic_base(const __atomic_base&) = delete;
>     __atomic_base& operator=(const __atomic_base&) = delete;
>     __atomic_base& operator=(const __atomic_base&) volatile = delete;
>
> In other words, libc++'s own std::atomic is supposed to be initialised by
> braces with a deleted copy constructor and assignment operator. It seems
> to me
> that it's a compiler bug.
>
> [1] http://llvm.org/svn/llvm-project/libcxx/trunk/include/atomic
> --
> Thiago Macieira - thiago.macieira (AT) intel.com
>   Software Architect - Intel Open Source Technology Center
>

Can you point out where in the standard brace assignment initialization is
not supposed to invoke a copy constructor? Because from what I've pointed,
I only see this case, and so I see both the compiler not working as
expected as both the qt and all this library code as ill-formed.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.qt-project.org/pipermail/development/attachments/20121203/877109aa/attachment.html>