[Bug c++/51283] g++ yields different behaviour for postfix increment and decrement operators with fundamental types vs. classes

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51283

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |INVALID

--- Comment #1 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-11-23 13:00:06 UTC ---
(In reply to comment #0)
>   int i = 0;
>   assert(i++ == i++); // OK

Undefined behaviour, try using -Wsequence-point

>   int* p = 0;
>   assert(p++ == p++); // OK

Undefined behaviour, try using -Wsequence-point

>   vector<int> v(10, 10);
>   vector<int>::iterator it = v.begin();
>   assert(it++ == it++); // FAILS

If vector::iterator is a pointer that's undefined behaviour.
If it's a class type (as in libstdc++) then the assertion is meant to fail
because it's equivalent to:

auto it1 = it++;
auto it2 = it++;
assert( it1 == it2 );  // obviously wrong

> GCC should calculate expression (it++ == it++) with an object copy returned by
> the overload it.operator ++ (int) method like for the first two examples.

No it shouldn't.