[Bug c++/62115] [5 Regression] ICE with invalid default argument

["roman.perepelitsa at gmail dot com" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62115

Roman Perepelitsa <roman.perepelitsa at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |roman.perepelitsa at gmail dot com

--- Comment #4 from Roman Perepelitsa <roman.perepelitsa at gmail dot com> ---
r216124 has also fixed another bug, which may result in valid code being
rejected and also in incorrect code generation for valid code.

#include <utility>

struct Base { int value; };
struct Derived : Base {};

void Foo(int&&) {}

int main() {
  Derived d;
  Foo(std::move(d).value);
}

Before r216124 this program produced the following compile error:

   test.cc: In function 'int main()':
    test.cc:10:25: error: cannot bind 'int' lvalue to 'int&&'
       Foo(std::move(d).value);
                         ^
    test.cc:6:6: error:   initializing argument 1 of 'void Foo(int&&)'
    void Foo(int&&) {}

If function Foo() was overloaded for const int& and int&, or if it used perfect
forwarding, it would result in successful compilation and incorrect runtime
behaviour.

Would it be possible to backport the fix to gcc-4_9-branch?