forcing tail/sibling call optimization

[Jeffrey A Law]

  In message < 20001126234619.A52E534D80@nile.gnat.com >you write:
  > Note that if you want to make a warning message a formal requirement,
  > then you must provide a formal definition of what a warning message is,
  > and I do not think you want to get into that sort of thing :-)
  > 
  > I think there is real merit in providing fairly formal specifications
  > for GNU C enhancements for three reasons.
  > 
  > 1. The attempt to produce a FFS can often reveal difficulties that are
  > not apparent in a less formal approach.
I couldn't agree more.  The rather informal approach has led to numerous
problems over the years - even for those extensions which are considered
"well documented".  

jeff