On trunk (also seen in gcc-4.4.0), gcc doesn't compile this snippet:
========================================
enum class E { A };
bool foo() { return E::A == E::A; }
========================================
in C++0x mode (cc1plus -quiet -std=c++0x enumtest.cpp -o /tmp/enumtest.s),
giving this error:
----------------------------------------
enumtest.cpp: In function 'bool foo()':
enumtest.cpp:2: error: invalid operands of types 'E' and 'E' to binary
'operator=='
----------------------------------------
I think this snippet is valid, since my interpretation of N2857's
[over.built]/15:
For every T, where T is an enumeration type or pointer to effective
object type, there exist candidate operator functions of the form
bool operator<(T , T );
bool operator>(T , T );
bool operator<=(T , T );
bool operator>=(T , T );
bool operator==(T , T );
bool operator!=(T , T );
is that type E should have a built-in operator==.
This is confirmed by the expected error message when compiling this snippet:
========================================
enum class E { A };
enum class F { B };
bool foo() { return E::A == F::B; }
========================================
which is:
----------------------------------------
enumtest2.cpp: In function 'bool foo()':
enumtest2.cpp:3: error: no match for 'operator==' in '(E)0 == (F)0'
enumtest2.cpp:3: note: candidates are: operator==(F, F) <built-in>
enumtest2.cpp:3: note: operator==(E, E) <built-in>
enumtest2.cpp:3: note: operator==(int, int) <built-in>
----------------------------------------
This message says there is indeed a built-in 'operator==(E, E)'. Shouldn't it
be found when compiling the first snippet?
This creates failures in Boost regression tests for runners using gcc-4.4.0 in
C++0x mode (as scoped enums have been enabled on that configuration recently).
--
Summary: No built-in comparison operators for scoped enums
(C++0x)
Product: gcc
Version: 4.4.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c++
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: frabar666 at gmail dot com
GCC build triplet: i686-pc-linux-gnu
GCC host triplet: i686-pc-linux-gnu
GCC target triplet: i686-pc-linux-gnu
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40193
