Re: [boost] [optional] Specializing optional to save space

On 27 September 2015 at 16:19, Andrey Semashev wrote:

On 27.09.2015 23:38, Nevin Liber wrote:

...

On 27 September 2015 at 15:03, David Stone wrote:

Andrzej posted another thread about creating a new compact_optional class.

...

The goal of this class is to make it easy to create a new optional type with a special sentinel value. This allows easy customization per instance of an optional value.

This post is about the opposite approach: making it easier to specialize optional to provide special behavior for all instances of a given optional type.

When I bring this up, people often compare it to the vector<bool> specialization, but this is a completely different issue.

No, it's the same issue. It'll have subtly different behavior. Generic programming counts on the behavior being the same.

The only reason vector<bool> is bad is because of its spelling. Had it been a different name, most people would be happy.

optional<T> controls the lifetime of T. If you have a sentinel instead, then the lifetime of the engaged T is exactly the same as the lifetime of optional<T>.

I got the impression that the OP intended that the specialized optional<> still controlled the lifetime of the adopted object the way it does now. The proposed change basically offloads engagement checking to a user-specializable trait but really nothing more than that.

His example was to use a specific value of std::string::size() as a sentinel. How does that work with his specialized optional if the specialization is orthogonal to lifetime control? Feel free to assume string and optional<string> are friends, as well as a specific implementation of std::string. -- Nevin ":-)" Liber mailto:nevin@eviloverlord.com (847) 691-1404