From: Jens Maurer (Jens.Maurer_at_[hidden])
Date: 2001-05-24 17:52:37
Darin Adler wrote:
> Well, this is only my opinion (and apparently Andrei's too, given what he
> said in his book). I'll try to be clear here. I'm not trying to win a
> debate, but just present my thinking in as transparent a way as possible.
I'm also of the opinion that having std::less for a smart pointer
is (and should be provided), but not operator<.
> That's why I added the std::less overload (not
> specialization, I guess) for shared_ptr.
std::less is a class, so it can't be an overload, only a
specialization.
> Whether shared_ptr's can also be
> compared in other contexts seemed less important to me,
It's undefined behavior to compare ordinary pointers with
operator<, unless they point to element of the same array.
Two shared_ptrs can never point into the same array, so
by analogy, there is never a reason to compare shared_ptr with
operator<, because the equivalent ordinary pointers cannot
be (legally) compared by operator< either. It's nice that
a shared_ptr class gives us the option to actually flag this
invalid usage as an error.
Jens Maurer