egcs, RFC: patch for possible -Wmissing-noreturn warning

[Jeffrey A Law]

  In message < 19981012095243.07133@albireo.ucw.cz >you write:
  >    It is not as simple as it might look on the first sight. It's easy to
  > prove there exists _no_ algorithm distinguishing functions that really return.
True, but if it is the compilation of the function itself (not the copy that
gets integrated into callers), then for this example we should be able to
determine that each and every path through the (trivial) function exits.

The other question is for a static inline function, we may halt the compilation
too early and as a result the things Kaveh wants to check hevn't been
initialized.

I'm well aware of the halting problem :-)

jeff