ZZ>
  >                                                    But how
  > do you distinguish a K&R compiler from an ANSI compiler if
  > not by testing for __STDC__ ?
ZZ>

  One of the ironies of life is that a non-ISO conformant C compiler is
  perfectly entitled to define __STDC__, and even to define it with the
  value 1.  So testing __STDC__ is *not* a valid test for whether a
  compiler is a conformant ISO C implementation or not.

  Fortunately, all of the C/C++ compilers for OS/2 provide all of the
  ISO C standard headers, and support prototypes and `void'.  There
  should be *no* reason for conditional compilation to cope with K&R C
  constructs.

  If people are still writing code that conditionally compiles for K&R
  C, then they are wasting their time needlessly.  The lowest common
  denominator for any platform is ISO C.  For OS/2, I don't think that
  any pre-ISO C compilers are even available.

ZZ>
  >                        But tell that to the zillion people
  > writing free software mostly for the unix world :)
ZZ>

  Incidentally, the FSF software is not the most well written software
  in the world.  I've studied some of the sources, and some of the
  programs are *terribly* badly written.

  For certain you should not be using the FSF sources as an ISO C
  validation suite.

  > JdeBP <
___
 X MegaMail 2.10 #0:
--- Maximus/2 3.00