Re: [boost] BOOST_NOEXCEPT macro?

Subject: Re: [boost] BOOST_NOEXCEPT macro?
From: Dave Abrahams (dave_at_[hidden])
Date: 2011-02-17 20:18:00

On Thu, Feb 17, 2011 at 7:02 PM, Emil Dotchevski
<emildotchevski_at_[hidden]> wrote:
> On Thu, Feb 17, 2011 at 12:51 PM, Dave Abrahams <dave_at_[hidden]> wrote:
>> On Thu, Feb 17, 2011 at 2:55 PM, Nevin Liber <nevin_at_[hidden]> wrote:
>>> On 17 February 2011 11:45, Dave Abrahams <dave_at_[hidden]> wrote:
>>>
>>>> On Thu, Feb 17, 2011 at 11:36 AM, Alexander Fokin <apfokin_at_[hidden]>
>>>> wrote:
>>>> > I believe it's not that simple.
>>>> >
>>>> > As I remember, throw() means the same thing as nothrow only in MSVC.
>>>> > For example, under GCC you'll have to use __attribute__((nothrow)).
>>>> >
>>>> > Correct me if I'm wrong.
>>>>
>>>> You're right.  In general, throw() has runtime semantics and can
>>>> produce runtime overhead in the form of increased executable size and
>>>> occasionally reduced speed.  noexcept has only compile-time semantics.
>>>>
>>>
>>> noexcept has runtime semantics as well, doesn't it?  If you are in a
>>> function declared noexcept(true) and an exception tries to escape,
>>> std::terminate is called.
>>
>> Grrr!  Right, I forgot they added that behavior against my advice.
>
> Do I understand correctly that they replaced the lame throw() standard
> behavior, to call unexpected() if an exception attempts to propagate
> out of the function, with calling std::terminate() instead? Am I
> missing something, isn't this equally lame?

Don't get me started :-(

-- 
Dave Abrahams
BoostPro Computing
http://www.boostpro.com