------- Comment #1 from pinskia at gcc dot gnu dot org 2008-12-12 23:04 -------
Why do you think __m could even appear in a static_assert? It is a function
argument. Changing __m to a type of int shows the same.
Changing it to a template argument causes an ICE though:
t.cc: In function 'int f()':
t.cc:11: internal compiler error: in tsubst_decl, at cp/pt.c:8622
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.
--- CUT ---
Testcase:
#include <cassert>
/// Enumeration for memory_order
typedef enum memory_order
{
memory_order_relaxed,
memory_order_consume,
memory_order_acquire,
memory_order_release,
memory_order_acq_rel,
memory_order_seq_cst
} memory_order;
class a
{
void
store1(void* __v, memory_order __m = memory_order_seq_cst) volatile
{
assert(__m == memory_order_acquire);
assert(__m == memory_order_acq_rel);
assert(__m == memory_order_consume);
}
template<memory_order __m = memory_order_seq_cst>
void
store2(void* __v)
{
static_assert(__m == memory_order_acquire,
"atomic_address::store requirements not met.");
static_assert(__m == memory_order_acq_rel,
"atomic_address::store requirements not met.");
static_assert(__m == memory_order_consume,
"atomic_address::store requirements not met.");
}
};
int f(void)
{
a b;
b.store2(0);
}
--
pinskia at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords|rejects-valid |ice-on-valid-code
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38502
