egcs, REVIEW FOR SUBMISSION, vfprintf.c ...

[Jeffrey A Law]

  In message < 199803312140.QAA28085@caip.rutgers.edu >you write:
  > 	Here is my first cut at a vfprintf.c for platforms with
  > stdarg/varargs but no vfprintf() function.  It fakes things by parsing
  > the format specifier string, pulling args and printing using fprintf.
  > I hacked it up using parts of libiberty/vasprintf.c and some RTFM.
  > 
  > 	Its currently meant to live in the gcc directory of egcs since
  > I'm not ready to tackle libiberty.  But it could easily live in the
  > libiberty directory with only a couple of changes in whcih include
  > files it gets.
  > 
  > 	I'd like it if people more familiar with the exact conformant
  > behavior of vfprintf() would look at it to make sure I got it right.
  > I'm especially unsure of the 'h' flag's meaning.
Wow.

Probably more complete than I would have done since systems without
vfprintf are diminishing regularly.

Even with the 'h' issue pending, I'd say go ahead and cobble up the
changes so that it can be installed -- it's a lot better than the
stuff we're currently using.

I suspect this will be linked into both the compiler (gcc, cpp,
cc1, etc) & the generator programs (gen*).  If so, you need
to handle it like rtl.o/HOST_RTL.





jeff