[PATCH 7/9] dix: Remove a bizzare path where we'd treat some m68k as sparc

Nix nix at esperi.org.uk
Thu May 22 16:38:00 PDT 2014


On 21 May 2014, Adam Jackson verbalised:

> Just what.
>
> Signed-off-by: Adam Jackson <ajax at redhat.com>
> ---
>  include/servermd.h | 11 -----------
>  1 file changed, 11 deletions(-)
>
> diff --git a/include/servermd.h b/include/servermd.h
> index 7f121c0..8d9c24f 100644
> --- a/include/servermd.h
> +++ b/include/servermd.h
> @@ -109,16 +109,6 @@ SOFTWARE.
>  
>  #endif                          /* __m32r__ */
>  
> -#if (defined(sun) && (defined(__sparc) || defined(sparc))) || \
> -    (defined(__uxp__) && (defined(sparc) || defined(mc68000))) || \
> -    defined(__sparc__) || defined(__mc68000__)
> -
> -#if defined(__sparc) || defined(__sparc__)
> -#if !defined(sparc)
> -#define sparc 1
> -#endif
> -#endif

I'm fairly sure this is not treating some m68k as sparc, but rather
checking for both sparc-based and ancient m68k suns, and arranging to
handle tiresome compiler variations where sparc-based suns could define
either __sparc or __sparc__ by defining 'sparc' unconditionally for
both.

Because having #defines with unadorned lower-case names like 'sparc' is
just a generally good idea and is not likely to cause problems at *all*.

So it's still horrible, but at least it's comprehensible.

-- 
NULL && (void)


More information about the xorg-devel mailing list