PRODUCT  :  Borland C++                           NUMBER  :  1018
  VERSION  :  3.1
       OS  :  WIN
     DATE  :  October 19, 1993                         PAGE  :  1/2

    TITLE  :  Address of Member function syntax using v3.1





  The following code highlights a change that was made to the
  Borland C++ v 3.1 compiler to allow for syntax used by the
  Microsoft Foundation Classes (MFC).  The issue revolves around
  the proper syntax used when taking the address of a member
  function.

  The four commented lines in the source below demonstrate the
  syntax in question.  The fourth example, &A::f, is the syntax
  that should be used for ANSI C++ compliance.  Using the option
  OPTIONS|COMPILER|SOURCE|ANSI, or -A at the command line, will
  generate the appropriate error when the &f or A::f syntax is
  used.

  The ANSI specification for C++ is currently still in draft form
  so this issue is subject to change.  Please accept our apologies
  for any inconvenience that this compiler modification has caused.


  class A
  {
  public:
      void f() {}
  };

  typedef void (A::*AFP)();

  class B : public A
  {
  public:
      static AFP afp1;
      static AFP afp2;
      static AFP afp3;
  };

  AFP afp4 = A::f;            // Fails when using ANSI C++
  AFP B::afp1 = &f;           // Fails when using ANSI C++
  AFP B::afp2 = A::f;         // Fails when using ANSI C++
  AFP B::afp3 = &A::f;        // Correct ANSI C++ syntax

  int main()
  {
      return 0;













  PRODUCT  :  Borland C++                           NUMBER  :  1018
  VERSION  :  3.1
       OS  :  WIN
     DATE  :  October 19, 1993                         PAGE  :  2/2

    TITLE  :  Address of Member function syntax using v3.1.




  }




  DISCLAIMER: You have the right to use this technical information
  subject to the terms of the No-Nonsense License Statement that
  you received with the Borland product to which this information
  pertains.