Re: [boost] [type_traits] A bug in has_nothrow_default_constructor

Subject: Re: [boost] [type_traits] A bug in has_nothrow_default_constructor
From: Andrzej Krzemienski (akrzemi1_at_[hidden])
Date: 2015-05-20 14:47:24

2015-05-20 19:00 GMT+02:00 John Maddock <jz.maddock_at_[hidden]>:

>
>
> On 20/05/2015 08:39, Andrzej Krzemienski wrote:
>
>> Hi All,
>> I have submitted a new ticket:
>> https://svn.boost.org/trac/boost/ticket/11324
>>
>> Given the information floating around that Boost.TypeTraits are
>> unmaintained and that they are to be replaced with the rewrite, is there
>> any chance that this is fixed, or at least the test case added to
>> regression matrix?
>>
> They're not unmaintained.  However, because of the risk of changes
> breaking a lot of stuff, we do try to keep the code ultra-stable.
> "Version2" is perhaps misnamed, as it's a more of a radical face-lift than
> a replacement.  Basically remove all the no-longer needed workarounds and
> simplify as much as possible.
>
>> BTW, does type_traits v2 handle this case correctly?
>>
> There should be no difference in behaviour between the two versions.
>
> BTW your test case works fine for me with MSVC and GCC, so I assume this
> is an Oracle compiler issue?
>

Yes, this is only on Oracle Solaris Studio C++ compiler (under
development). The developers say it is a known issue with this compiler
that they intend to fix. What bugs me is that tests for Boost.Optional fail
because of this reason, but tests in TypeTraits all pass.

Regards,
&rzej