Re: killing STDC_0_IN_SYSTEM_HEADERS

[Jeffrey A Law <law at cygnus dot com>]

  In message <20000107221111.A2212@wolery.cumb.org>you write:
  > > I originally installed the __STDC__=0 solution, but RMS felt that it
  > > was better for user programs if __STDC__ was 1 on all platforms
  > > (unless you specify -traditional, of course).  This makes it easier to
  > > port user programs that assume GCC semantics.
  > > 
  > > In other words, he felt it more important to promote compatibility
  > > among GCC platforms than to promote compatibility between GCC and Sun C.
  > > Both goals are important, but here they're mutually exclusive.
  > 
  > RMS has a valid point, but I think the larger goal - encouraging people
  > to write portable code - would be better served by matching the vendor
  > compiler.  Since Solaris (and other SVR4 compilers, apparently) do define
  > __STDC__ to 0 in a user visible fashion, it is not portable to test __STDC_
Sorry, no.

Consistency across GCC implementations has almost always had a higher priority
than consistency with a vendor compiler (the most obvious exception is the
debugging info stuff).

You'll need more than a simple statement to convince us to change directions
on this issue.

jeff