Re: [boost] is_single_element?

From: Damien Fisher (damien_at_[hidden])
Date: 2002-05-15 21:02:55

On Wed, 15 May 2002, Dave Harris wrote:

> In-Reply-To: <Pine.GSO.4.05.10205160011080.5723-100000_at_galois>
> On Thu, 16 May 2002 00:18:41 +1000 (EST) Damien Fisher 
> (damien_at_[hidden]) wrote:
> > I would like to see something like this in Boost:
> > 
> > template<typename Container>
> > bool is_single_element(const Container &cont);
> 
> Should it be defined for iterator pairs as well as, or instead of, 
> containers? Most algorithms work on iterators.

I like this.  Someone pointed out to me it in a private e-mail that a
length-checked distance function might fit in better with the current
STL naming, eg:

template<typename FwdIt>
std::iterator_traits<FwdIt>::distance_type n
distance_n(
    FwdIt first, FwdIt last,
    std::iterator_traits<FwdIt>::distance_type n
)
{
    std::iterator_traits<FwdIt>::distance_type i = 0;
    while(i != n && first != last)
    {
        ++first;
        ++i;
    }

    return i;
}

With an appropriate specialization for random-access iterators.

Do people prefer this?  If so, I can do it properly (add the appropriate
specialization), add test cases and docs, etc.  I have wanted this many
times, and it seems to have gotten a reasonable response (mostly
pointing out the mistakes in my example implementation, which I hadn't
really meant as a real implementation, and which was written late at
night, but, hey, mea culpa for writing e-mails half-asleep :) ).