[llvm-bugs] [Bug 34009] New: Clang 5.0 consider template method unevaluated and accepts this code, GCC rejects it

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=34009

            Bug ID: 34009
           Summary: Clang 5.0 consider template method unevaluated and
                    accepts this code, GCC rejects it
           Product: clang
           Version: trunk
          Hardware: PC
                OS: All
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: C++11
          Assignee: unassignedclangbugs at nondot.org
          Reporter: joker.eph at gmail.com
                CC: dgregor at apple.com, llvm-bugs at lists.llvm.org

Clang accepts the following (C++11) code while GCC rejects it (4.9 to 7.0).
Clang seems to consider that the constexpr `IS_N` is enough to make the method
`bar()` unevaluated, while gcc complains. I don't understand what makes clang
happy here?

Changing the define makes both compiler happy, as the constexpr is dependent on
the template args.



#include <type_traits>

#define INDEPENDENT_CONSTEXPR 1

template<int N = 1>
struct Bar {
  typename std::enable_if<N != 1, void>::type foo() {
  }
};

template<int N = 1>
struct Foo {
#if INDEPENDENT_CONSTEXPR
    static constexpr bool IS_N = true;
#else 
    static constexpr bool IS_N = (N == 1);
#endif
    // (N == 1);
    using myBar = Bar<IS_N>;

    void bar() {
        myBar b;
        b.foo();
    }
};


void foobar() {
  Foo<0> f0;
  Foo<1> f1;
#if !INDEPENDENT_CONSTEXPR
  f0.bar();
#endif
// Following should never work
//  f1.bar();
}

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20170731/111c623a/attachment-0001.html>