I think that the name of the algorithm should contain the word «partition». Something like partition_by_unique. Of course the algorithm also must be defined with a binary predicate.

I have one more problem with algorithm names. Early I suggested an algorithm that checks that a sequence is a strictly sorted that is that there is no equal elements in a sorted sequence. The standard algorithm std::is_sorted or std::is_sorted_until is unable to perform such a check. Could such an algorithm be named like is_strictly_sorted or a better name can be proposed?

You can meet me at http://cpp.forum24.ru/ or www.stackoverflow.com or http://ru.stackoverflow.com

On 12/01/2024 19:03, Vladimir Grigoriev wrote:

> I am sure that it is a drawback of the library of
> the standard algorithms. What alternative can you suggest?

I wouldn't call it a "drawback"; it's a *design* of the std::unique
algorithm. You're proposing a variation of it that doesn't eliminate the
equivalent elements in each run, it just moves them. That's completely
fine to propose if you have a use case for it! Just find a good name for
this variation (which is really the only thing I've been questioning),
properly define the semantics of what happens with the elements after
the first of each run, and create a suitable proposal.

My 2 c,
--
Giuseppe D'Angelo