Subject: Re: xdr_float.c?
To: Gordon W. Ross <>
From: Aaron J. Grier <>
List: tech-toolchain
Date: 03/31/1999 09:59:43
On Wed, Mar 31, 1999 at 02:23:10PM +1000, Simon Burge wrote:
> "Gordon W. Ross" wrote:
> > How about the trivial solution below?
> > 
> > Add this somewhere widely included:
> > i.e. src/sys/arch/mips/include/cdefs.h
> > 
> > #if defined(__mipseb__) || defined(__mipsel_)
> > #define __mips__ 1
> > #endif

This doesn't fix lint, since lint only checks MACHINE_ARCH, and not the
definition of the mipse{l,b}.

> I'm tempted (as Soda suggests) to ``fix'' lint - after all, it's not
> a deficiency in the toolchain - it gets everything right.

Yeah, it's basically lint that's the problem...

However, this still raises the interesing question of whether endian-ness
even belongs in MACHINE_ARCH.  Obviously for some processors endian-ness
is determined by MACHINE_ARCH, so besides Mips and Hitachi, what other
processors can run in big or little endian modes?

  Aaron J. Grier  | "Not your ordinary poofy goof." |
        I do not support US and NATO "police actions" in Yugoslavia.