Re: C++: constant folding in a conditional expr

[Andrew Haley <aph at cambridge dot redhat dot com>]

Nathan Sidwell writes:
 > Andrew Haley wrote:
 > > This test case fails if BROKEN is #defined:
 > This test case is illformed, regardless of BROKEN's #definedness :-)
 > 
 > [9.4.2]/4 requires an out of class definition of static data members
 > yea, even if the in-class declaration is of a constant type and
 > contains an initializer.

Right.  I see

"If a static data member is of const integral or const enumeration
type, its declaration in the class definition can specify a
constant-initializer which sahll be an integral constant expression.
In that case, a member can appear in integral constant expressions
within its scope."

So, what would the legal form of this program be?  Like this, I guess:

class Foo
{
  public:
    static const unsigned ACONST = 0;
    static const unsigned BCONST = 1;

    void SetData(bool p);

  private:
    unsigned _data;
};

const unsigned Foo::ACONST;
const unsigned Foo::BCONST;

 > > I've written a patch that doesn't treat the conditional expr as an
 > > lvalue if either arm is read only, and that seems to fix the problem,
 > > but is it the right fix?
 > I don't think so. What if I wrote
 > 	ptr = &(data ? ACONST : BCONST);

Ah.  :-)

Thanks very much,
Andrew.