Re: Fixed String review

From: Krystian Stasiowski (sdkrystian_at_[hidden])
Date: 2019-11-27 04:55:59

>   In the same vein, more of it should be noexcept - why aren't
empty() and size() ?

empty() and size() not being noexcept is more of an oversight on our part.
As for why more of the member functions aren't noexcept, they have the same
exception specifications as those of their counterparts in std::string.
That being said, although they do call "dangerous" operations like
Traits::length that are not noexcept, such functions don't throw (at least
for std::char_traits), so its difficult to specify as noexcept since Traits
could also be a user defined type.

>  You can add a specialisation for size 0 that turns it into an empty
struct.

Certainly, it would be fairly trivial to do so.

> I suggested changing this to the minimum size needed to store values 0 to
N in the previous email discussion, but that hasn't been implemented.

Sure, that's extremely trivial to do, and seems like a great idea.

On Tue, Nov 26, 2019 at 7:25 PM Peter Dimov via Boost <boost_at_[hidden]>
wrote:

> Gavin Lambert wrote:
> > On 27/11/2019 06:46, Peter Dimov wrote:
> > > Vinnie Falco wrote:
> > >
> > >> In which use-cases do you anticipate the need for a constexpr
> > >> static_string?
> > >
> > > When manipulating strings at compile time? Such as s1 += s2?
> >
> > Wouldn't that use case be better served with an immutable string and
> > operator+?
>
> Not in general.
>
> constexpr string_view f( int x );
>
> constexpr auto g( int n )
> {
>     fixed_string<512> s;
>
>     for( int i = 0; i < n; ++i )
>     {
>         s += f( i );
>     }
>
>     return s;
> }
>
> You can rewrite this specific example using recursion but in general, it'd
> be painful and unreadable.
>
>
> _______________________________________________
> Unsubscribe & other changes:
> http://lists.boost.org/mailman/listinfo.cgi/boost
>