https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90857
Jonathan Wakely <redi at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |NEW
Last reconfirmed| |2019-06-12
Resolution|INVALID |---
Ever confirmed|0 |1
--- Comment #2 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to Jonathan Wakely from comment #1)
> (In reply to Jose Dapena Paz from comment #0)
> > As far as I know, the specification does not really say this is a valid
> > case.
>
> It explicitly says it's NOT valid:
>
> Expects: The iterator following position is dereferenceable.
>
Sorry, I copied the wrong paragraph. The relevant one is:
Expects: All iterators in the range (position, last) are dereferenceable.
The range (position, last) is a closed interval, which we've interpreted to
mean that the range [position+1,last) must be valid. But I see your point,
that it could also be interpreted to mean that pos==last is a no-op.
I'll confirm this and consider it further.
