Re: [boost] [optional] Safe optional

Subject: Re: [boost] [optional] Safe optional
From: Andrey Semashev (andrey.semashev_at_[hidden])
Date: 2014-11-18 15:47:20

On Tuesday 18 November 2014 13:39:05 Felix Uhl wrote:
> Matt Calabrese wrote:
> 
> 
> 
> > I know that there are others (Eric?) who prefer to think of optional as a
> > container of size 0 or 1, so it could also make sense to provide an
> > accessor to a range. One interesting effect of this is that now
> > range-based-for sort of becomes a "visit if." There are probably other...
> > interesting uses of standard algorithms.
> 
> 
> 
> I don’t know why, but I really like that idea;
> 
> 
> int f(optional<int> input){
> 
>     int ret = 0;
> 
>     for(int opt : input){
> 
>         ret = opt;
> 
>     }
> 
>     return ret;
> 
> }

Looks weird to me. It will look unintuitive to newcomers, IMHO. I remember it 
was always strange to me that to clear an optional I had to assign none 
instead of calling clear(). This is the same kind of thing.

> It somehow seems like a really interesting Syntax, but to be honest,
> 
> it doesn’t really offer any benefits, does it?
> 
> You neither have any improvements on performance, nor any on 
> 
> code clarity.

The performance can be worse than a traditional if. The for statement requires 
iterator increment, comparison and a conditional jump. That is in addition to 
the test if the optional is empty and returning pseudo-iterators. Compilers 
may optimize this overhead, though.