[Xcb] Re: CRAY bitfield support in protocol headers: does anyone care?

[Juliusz Chroboczek]

I think this is getting pretty much irrelevant for this list.  If it's
too much bother to keep the Cray-specific macros, I suggest we just
delete them, and hope that either nobody cares about Crays any more,
or Cray have fixed their compiler.  (I'm willing to bet a beer on the
former.)

Should this assumption prove wrong, it should be easy enough to revert
the changes.

However, since you're asking,

> Doesn't POSIX actually provide fixed size types?

I believe it's C99 (Section 7.18.1.1), not POSIX directly (although
POSIX.2001 requires C99).  However, an implementation does not need to
implement them -- the spec still allows implementing C on the PDP-20
and on machines with a sign-magnitude representation of integers:

  These types are optional. However, if an implementation provides
  integer types with widths of 8, 16, 32, or 64 bits, no padding bits,
  and (for the signed types) that have a two's complement representation,
  it shall define the corresponding typedef names.

                                        Juliusz