Created on 2013-02-12.00:00:00 last changed 95 months ago
[Moved to DR at the November, 2014 meeting.]
Proposed resolution (June, 2013):
This issue is resolved by the proposed resolution of issue 1351.
The current specification is not clear whether the exception-specification for a function is propagated to the result of taking its address. For example:
template<class T> struct A { void f() noexcept(false) {} void g() noexcept(true) {} }; int main() { if (noexcept((A<short>().*(&A<short>::f))())) return 1; if (!noexcept((A<long>().*(&A<long>::g))())) return 1; return 0; }
There is implementation variance on whether main returns 0 or 1 for this example. (It also appears that taking the address of a member function of a class template requires instantiating its exception-specification, but it is not clear whether the Standard currently specifies this or not.)
History | |||
---|---|---|---|
Date | User | Action | Args |
2017-02-06 00:00:00 | admin | set | status: drwp -> cd4 |
2015-05-25 00:00:00 | admin | set | status: dr -> drwp |
2015-04-13 00:00:00 | admin | set | messages: + msg5425 |
2014-11-24 00:00:00 | admin | set | status: ready -> dr |
2014-07-07 00:00:00 | admin | set | status: review -> ready |
2014-05-27 00:00:00 | admin | set | status: ready -> review |
2014-03-03 00:00:00 | admin | set | status: review -> ready |
2013-05-03 00:00:00 | admin | set | messages: + msg4327 |
2013-05-03 00:00:00 | admin | set | status: open -> review |
2013-02-12 00:00:00 | admin | create |