On 16 October 2013 10:53, Gabriel Dos Reis <gdr@axiomatics.org> wrote:
Ville Voutilainen <ville.voutilainen@gmail.com> writes:

| On 16 October 2013 18:46, Gabriel Dos Reis <gdr@axiomatics.org> wrote:
|
|
|     |     My own personal view (not that of chair) is that if std::less<T>(l,r)
|     and
|     |     "l < r" are
|     |     both defined, then they should yield the same answer.
|     |
|     |
|     | Which fails for pointers.
|
|     "fails" in which sense?  It is certainly true in the current standards
|     these expressions are both defined when 'l' and 'r' are related
|     addresses (relative to the same object), which was exactly my point.
|
| Fails in the sense that less<int*>(l, r) and l<r do not necessarily yield the
| same answer.

Only on the cases, e.g. input values, where they aren't both defined.

Only because they have different preconditions.  They shouldn't.
--
 Nevin ":-)" Liber  <mailto:nevin@eviloverlord.com>  (847) 691-1404