Re: [boost] variant2 - really dumb way off topic question here

Subject: Re: [boost] variant2 - really dumb way off topic question here
From: Andrzej Krzemienski (akrzemi1_at_[hidden])
Date: 2019-02-28 13:29:40

sob., 23 lut 2019 o 01:12 Robert Ramey via Boost <boost_at_[hidden]>
napisał(a):

> On 2/22/19 9:48 AM, Peter Dimov via Boost wrote:
> > I'd like to request a formal Boost review for the Variant2 library,
> > https://github.com/pdimov/variant2. Variant2 contains an implementation
> > of a
> > variant<T...> type that is an almost conforming std::variant, except it
> > doesn't require C++17 and supports C++11 and above.
> >
> > It's also never valueless and has a few other extensions.
>
> I see optional, expected, outcome and .. (monad?) as just special cases
> of variant. for example
>
> template<typename T>
> using optional = variant<false_type, T>;
>
> Why is it necessary to have all these type implemented?  Can't there be
> some sort of "base" type which can be used to implement all these others?
>

I think it is about the interfaces and guarantees. optional provides
.value_or(), .map() functions, and operators, like ->. Also, because it
does not have to consider the assignment of mixed types its implementation
is more lightweight and therefore compiles faster and introduces fewer
symbols.

Regards,
&rzej;