Subject: Re: [boost] [variant] Basic rvalue and C++11 features support
From: Peter Dimov (lists_at_[hidden])
Date: 2013-01-08 14:03:09
Nevin Liber wrote:
> On 8 January 2013 12:09, Peter Dimov <lists_at_[hidden]> wrote:
> > Yes, I agree. To expand on that, the requirements defined by the
> > standard
> > library such as CopyConstructible, LessThanComparable, and so on, do not
> > have preconditions. So if a movable type claims to conform to these
> > requirements, it can't impose a precondition that the operations only
> > work
> > on values that haven't been moved from.
>
> I believe it can. Take n3485 25.4p3 (Sorted and related operations):
> "For algorithms other than those described in 25.4.3 to work correctly,
> comp has to induce a strict weak ordering on the values."
>
> This, for instance, allows sorting of doubles as long as none of them are
> NaN.
If I have the values { 1, 7, 5, -3 } and a predicate that works correctly
for these specific values, but returns garbage for others, does it satisfy
the requirement? An interesting question... for another time, maybe.